{"generatedBy":"scripts/modelChanges.ts","what":"Changes to the config, tokenizer, chat template, generation defaults or weights of tracked open-weight models on the Hugging Face Hub, one row per commit.","headline":"238 of 409 tracked open-weight models changed a tracked key after their first upload; 124 changed a structural one, 69 of them more than a week after release.","stats":{"models":409,"snapshots":1699,"changes":586,"byClass":{"structural":188,"tokenizer":185,"template":160,"generation-defaults":29,"weights-only":0,"metadata":24},"modelsWithStructural":124,"modelsWithAnyChange":238,"earliest":"2019-02-18T10:36:16.000Z","lastFetched":"2026-10-18T09:00:52.565Z","daily":0,"afterFirstWeek":313,"modelsWithStructuralAfterFirstWeek":69},"rawData":"data/model-changes/snapshots.jsonl and changes.jsonl in the Neurarch repository; this file carries every change row.","changes":[{"id":"zai-org/GLM-5.3","date":"2026-10-14T12:22:48.000Z","classes":["template"],"sentence":"GLM-5.3: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2026-10-14).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"3740abcea51c4583","after":"0e2b6268ac54f04d"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"aca966e4e02791568aa6a4ced368624b3d897f42","toSha":"ba32f8c19c8738e2c57c71798c601d477a7bf844","commit":"https://huggingface.co/zai-org/GLM-5.3/commit/ba32f8c19c8738e2c57c71798c601d477a7bf844","source":"backfill"},{"id":"baidu/Unlimited-OCR","date":"2026-10-14T06:17:59.000Z","classes":["structural","template"],"sentence":"Unlimited-OCR: model_type \"unlimited-ocr\" to \"unlimited_ocr\"; a different class is instantiated from the same id; a chat template was added; apply_chat_template now formats prompts that it previously refused or formatted by default (2026-10-14).","keys":[{"file":"config","key":"model_type","before":"unlimited-ocr","after":"unlimited_ocr"},{"file":"tokenizer","key":"chat_template_sha256","before":null,"after":"690dcec206a691b4"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"d549bb9d6a055dbe291408916d66acc2cd5920f6","toSha":"8249d20b2822a7d8a4c22c8d8c9069b5ac36bda3","commit":"https://huggingface.co/baidu/Unlimited-OCR/commit/8249d20b2822a7d8a4c22c8d8c9069b5ac36bda3","source":"backfill"},{"id":"unsloth/gpt-oss-20b-BF16","date":"2026-10-05T07:53:40.000Z","classes":["tokenizer","template"],"sentence":"gpt-oss-20b-BF16: EOS changed from [200002,199999] to [200002,199999,200012]; anything fine-tuned or served against the old value stops at a different token; the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2026-10-05).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"445c3a7c29d9cf61","after":"949f918f62563e2d"},{"file":"generation","key":"eos_token_id","before":[200002,199999],"after":[200002,199999,200012]}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":{"before":[200002,199999],"after":[200002,199999,200012]},"bos":null},"fromSha":"cc89b3e7fd423253264883a80a4fa5abc619649f","toSha":"5a6fe4c0c1a86147cb9dc1bb99486e61d12a3c9e","commit":"https://huggingface.co/unsloth/gpt-oss-20b-BF16/commit/5a6fe4c0c1a86147cb9dc1bb99486e61d12a3c9e","source":"backfill"},{"id":"openbmb/MiniCPM5-2B","date":"2026-09-29T06:12:33.000Z","classes":["template"],"sentence":"MiniCPM5-2B: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2026-09-29).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"cc945752db555d60","after":"bc1f78ef8a86b6bd"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"ab8fd77d37aa92791af5d0b4db99c73355fc14a2","toSha":"f97400052a43d642bbc6e9975e2397e3ae6a6b52","commit":"https://huggingface.co/openbmb/MiniCPM5-2B/commit/f97400052a43d642bbc6e9975e2397e3ae6a6b52","source":"backfill"},{"id":"biohub/ESMC-6B","date":"2026-09-14T19:45:56.000Z","classes":["structural","tokenizer"],"sentence":"ESMC-6B: EOS declared as 2 (was unset); generation that relied on the library default now stops on this token; BOS declared as 0 (was unset); head dim unset to 64, MLP width unset to 6912, attention heads unset to 40, layer count unset to 80, KV heads unset to 40 (2026-09-14).","keys":[{"file":"config","key":"architectures","before":["ESMCForMaskedLM"],"after":["EsmcForMaskedLM"]},{"file":"config","key":"bos_token_id","before":null,"after":0},{"file":"config","key":"eos_token_id","before":null,"after":2},{"file":"config","key":"head_dim","before":null,"after":64},{"file":"config","key":"intermediate_size","before":null,"after":6912},{"file":"config","key":"max_position_embeddings","before":null,"after":2048},{"file":"config","key":"num_attention_heads","before":null,"after":40},{"file":"config","key":"num_hidden_layers","before":null,"after":80},{"file":"config","key":"num_key_value_heads","before":null,"after":40},{"file":"config","key":"rope_theta","before":null,"after":10000},{"file":"tokenizer","key":"tokenizer_class","before":"ESMCTokenizer","after":"EsmcTokenizer"}],"consequence":{"params":{"before":3358597120,"after":4935700480},"weightsGbBf16":{"before":6.717,"after":9.871},"trainPeakGb":{"before":128.014,"after":456.812},"fingerprint":null,"context":{"before":null,"after":2048},"eos":{"before":null,"after":2},"bos":{"before":null,"after":0}},"fromSha":"2f13fa295be39898399cf7f6b0a97c26a13d5b3c","toSha":"7fe00682557aa7688b19a4dbcb21b267da99b7e6","commit":"https://huggingface.co/biohub/ESMC-6B/commit/7fe00682557aa7688b19a4dbcb21b267da99b7e6","source":"backfill"},{"id":"CMSManhattan/JiRackUltra_1b","date":"2026-09-05T15:24:21.000Z","classes":["template"],"sentence":"JiRackUltra_1b: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2026-09-05).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"56a1447ad31926fd","after":"f8db9a4e5f4bae1a"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"35af48943f310be3edce49361043f8bbc961fd4a","toSha":"9662105ae912c8d8c705cf7cef9645c049e89923","commit":"https://huggingface.co/CMSManhattan/JiRackUltra_1b/commit/9662105ae912c8d8c705cf7cef9645c049e89923","source":"backfill"},{"id":"zai-org/GLM-5.3","date":"2026-09-04T06:41:23.000Z","classes":["template"],"sentence":"GLM-5.3: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2026-09-04).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"69bb3ab52067898e","after":"3740abcea51c4583"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"7cda81930d6e4cef42f48555de830aa32ecdde28","toSha":"aca966e4e02791568aa6a4ced368624b3d897f42","commit":"https://huggingface.co/zai-org/GLM-5.3/commit/aca966e4e02791568aa6a4ced368624b3d897f42","source":"backfill"},{"id":"farbodtavakkoli/OTel-2.0-LLM-31B-IT","date":"2026-09-01T15:45:54.000Z","classes":["structural","tokenizer","template","metadata"],"sentence":"OTel-2.0-LLM-31B-IT: EOS changed from 106 to [1,106]; anything fine-tuned or served against the old value stops at a different token; derived parameters 30.15B to 30.56B (+418.7M, +0.84 GB of bf16 weights); architectures declared as [\"Gemma4ForConditionalGeneration\"] (was unset) (2026-09-01).","keys":[{"file":"config","key":"architectures","before":null,"after":["Gemma4ForConditionalGeneration"]},{"file":"config","key":"eos_token_id","before":106,"after":[1,106]},{"file":"config","key":"model_type","before":"gemma4_text","after":"gemma4"},{"file":"tokenizer","key":"chat_template_sha256","before":"ae53464bf3be2580","after":"7d35f613f292abfc"}],"consequence":{"params":{"before":30146156544,"after":30564872304},"weightsGbBf16":{"before":60.292,"after":61.13},"trainPeakGb":{"before":1630.22,"after":1745.918},"fingerprint":{"before":"5acf84f3","after":"79e55b50"},"context":null,"eos":{"before":106,"after":[1,106]},"bos":null},"fromSha":"84e7ead1824ba47765a33403f6eda19c77f34f43","toSha":"522937ff14a94c47b2d2960b95e474ec7ab40aa2","commit":"https://huggingface.co/farbodtavakkoli/OTel-2.0-LLM-31B-IT/commit/522937ff14a94c47b2d2960b95e474ec7ab40aa2","source":"backfill"},{"id":"ornith-ai/Ornith-1.5-397B-NVFP4","date":"2026-08-29T04:35:55.000Z","classes":["structural"],"sentence":"Ornith-1.5-397B-NVFP4: quantization_config changed; the stored format of the weights is different (2026-08-29).","keys":[{"file":"config","key":"quantization_config","before":{"quant_algo":"NVFP4","quant_method":"modelopt","_rest":"e5941c12dee63232"},"after":{"quant_algo":"MIXED_PRECISION","quant_method":"modelopt","_rest":"323fbb0514b5470b"}}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"69f1462c2069c8d427c459e58972334af9406fd6","toSha":"23a6070036a047432180c90b0e9c2837469bba96","commit":"https://huggingface.co/ornith-ai/Ornith-1.5-397B-NVFP4/commit/23a6070036a047432180c90b0e9c2837469bba96","source":"backfill"},{"id":"RadixArk/Qwen3.8-27B-DSpark","date":"2026-08-28T03:59:22.000Z","classes":["structural","tokenizer"],"sentence":"Qwen3.8-27B-DSpark: EOS changed from 248046 to 248044; anything fine-tuned or served against the old value stops at a different token; BOS 248044 removed; MLP width 10240 to 17408, attention heads 40 to 32 (2026-08-28).","keys":[{"file":"config","key":"bos_token_id","before":248044,"after":null},{"file":"config","key":"eos_token_id","before":248046,"after":248044},{"file":"config","key":"intermediate_size","before":10240,"after":17408},{"file":"config","key":"num_attention_heads","before":40,"after":32}],"consequence":{"params":{"before":3644106240,"after":4142179840},"weightsGbBf16":{"before":7.288,"after":8.284},"trainPeakGb":{"before":212.629,"after":239.278},"fingerprint":null,"context":null,"eos":{"before":248046,"after":248044},"bos":{"before":248044,"after":null}},"fromSha":"923ed3a8572615643f0137e424e4ce4edd7f1cda","toSha":"d0755f9c75414ab679101ea75453d7b4687a20d4","commit":"https://huggingface.co/RadixArk/Qwen3.8-27B-DSpark/commit/d0755f9c75414ab679101ea75453d7b4687a20d4","source":"backfill"},{"id":"ai-sage/Giga-Embeddings-instruct-480M-0826","date":"2026-08-24T13:23:12.000Z","classes":["structural"],"sentence":"Giga-Embeddings-instruct-480M-0826: context window 32768 to 8192 tokens (2026-08-24).","keys":[{"file":"config","key":"max_position_embeddings","before":32768,"after":8192}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":{"before":32768,"after":8192},"eos":null,"bos":null},"fromSha":"eb3d868e91bc83c6030b9722f6155ce89ffb0fc3","toSha":"2d0c1a92716eef0e5b6972df85b5883eb5b4f57a","commit":"https://huggingface.co/ai-sage/Giga-Embeddings-instruct-480M-0826/commit/2d0c1a92716eef0e5b6972df85b5883eb5b4f57a","source":"backfill"},{"id":"OBLITERATUS/Qwen3.8-27B-OBLITERATED","date":"2026-08-23T17:42:37.000Z","classes":["template"],"sentence":"Qwen3.8-27B-OBLITERATED: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2026-08-23).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"c02aa2b34e5ba89b","after":"4c2b3eabcdd87d5f"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"2648a6231b82328c601ba27b9ffd5029057d0e33","toSha":"489c6c867824d9111e7c79e1ff645e78491fa7e2","commit":"https://huggingface.co/OBLITERATUS/Qwen3.8-27B-OBLITERATED/commit/489c6c867824d9111e7c79e1ff645e78491fa7e2","source":"backfill"},{"id":"OBLITERATUS/Qwen3.8-27B-OBLITERATED","date":"2026-08-23T03:58:25.000Z","classes":["tokenizer"],"sentence":"Qwen3.8-27B-OBLITERATED: EOS changed from 248044 to 248046; anything fine-tuned or served against the old value stops at a different token; pad token unset to 248044; batched training masks a different id (2026-08-23).","keys":[{"file":"config","key":"eos_token_id","before":248044,"after":248046},{"file":"config","key":"pad_token_id","before":null,"after":248044}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":{"before":248044,"after":248046},"bos":null},"fromSha":"5350c98c83e37447f416ffa88f2982d9794f6b6a","toSha":"2648a6231b82328c601ba27b9ffd5029057d0e33","commit":"https://huggingface.co/OBLITERATUS/Qwen3.8-27B-OBLITERATED/commit/2648a6231b82328c601ba27b9ffd5029057d0e33","source":"backfill"},{"id":"OBLITERATUS/Qwen3.8-27B-OBLITERATED","date":"2026-08-23T03:58:21.000Z","classes":["tokenizer"],"sentence":"Qwen3.8-27B-OBLITERATED: EOS changed from [248046,248044] to [248046]; anything fine-tuned or served against the old value stops at a different token (2026-08-23).","keys":[{"file":"generation","key":"eos_token_id","before":[248046,248044],"after":[248046]}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":{"before":[248046,248044],"after":[248046]},"bos":null},"fromSha":"d39e846a298fe4bff3d749fed0b4a5ec998377f7","toSha":"5350c98c83e37447f416ffa88f2982d9794f6b6a","commit":"https://huggingface.co/OBLITERATUS/Qwen3.8-27B-OBLITERATED/commit/5350c98c83e37447f416ffa88f2982d9794f6b6a","source":"backfill"},{"id":"OBLITERATUS/Qwen3.8-27B-OBLITERATED","date":"2026-08-22T23:37:24.000Z","classes":["structural","tokenizer","template"],"sentence":"Qwen3.8-27B-OBLITERATED: derived parameters 26.79B to 26.79B (+7.0M, +0.01 GB of bf16 weights); architectures [\"Qwen3_5ForCausalLM\"] to [\"Qwen3_5ForConditionalGeneration\"]; a different class is instantiated from the same id; the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2026-08-22).","keys":[{"file":"config","key":"architectures","before":["Qwen3_5ForCausalLM"],"after":["Qwen3_5ForConditionalGeneration"]},{"file":"config","key":"model_type","before":"qwen3_5_text","after":"qwen3_5"},{"file":"tokenizer","key":"add_bos_token","before":null,"after":false},{"file":"tokenizer","key":"chat_template_sha256","before":"1bffd744ab18e116","after":"c02aa2b34e5ba89b"}],"consequence":{"params":{"before":26786168832,"after":26793185024},"weightsGbBf16":{"before":53.572,"after":53.586},"trainPeakGb":{"before":1557.235,"after":1657.854},"fingerprint":{"before":"40a1bb72","after":"023cc914"},"context":null,"eos":null,"bos":null},"fromSha":"73c3fd6ac5391f0259c8eeb39dcc796263f7e67e","toSha":"d39e846a298fe4bff3d749fed0b4a5ec998377f7","commit":"https://huggingface.co/OBLITERATUS/Qwen3.8-27B-OBLITERATED/commit/d39e846a298fe4bff3d749fed0b4a5ec998377f7","source":"backfill"},{"id":"OBLITERATUS/Qwen3.8-27B-OBLITERATED","date":"2026-08-22T08:57:29.000Z","classes":["structural","tokenizer"],"sentence":"Qwen3.8-27B-OBLITERATED: derived parameters 26.79B to 26.79B (-7.0M, -0.01 GB of bf16 weights); architectures [\"Qwen3_5ForConditionalGeneration\"] to [\"Qwen3_5ForCausalLM\"]; a different class is instantiated from the same id; add_bos_token false to unset; encoding the same text yields different ids (2026-08-22).","keys":[{"file":"config","key":"architectures","before":["Qwen3_5ForConditionalGeneration"],"after":["Qwen3_5ForCausalLM"]},{"file":"config","key":"model_type","before":"qwen3_5","after":"qwen3_5_text"},{"file":"tokenizer","key":"add_bos_token","before":false,"after":null}],"consequence":{"params":{"before":26793185024,"after":26786168832},"weightsGbBf16":{"before":53.586,"after":53.572},"trainPeakGb":{"before":1657.854,"after":1557.235},"fingerprint":{"before":"023cc914","after":"40a1bb72"},"context":null,"eos":null,"bos":null},"fromSha":"fc09b0df0896106d9bd6bcd7403905dd83fe12d7","toSha":"73c3fd6ac5391f0259c8eeb39dcc796263f7e67e","commit":"https://huggingface.co/OBLITERATUS/Qwen3.8-27B-OBLITERATED/commit/73c3fd6ac5391f0259c8eeb39dcc796263f7e67e","source":"backfill"},{"id":"bloomer010/Ling-3.0-tiny-GGUF","date":"2026-08-21T02:45:42.000Z","classes":["template"],"sentence":"Ling-3.0-tiny-GGUF: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2026-08-21).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"eb6226c94ae38058","after":"ef75b06f1e94e0ca"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"bc488bc165f53791586c04df99084859a0f0d59f","toSha":"4d07e0bfc17ea9370951584a54589a7452b2693f","commit":"https://huggingface.co/bloomer010/Ling-3.0-tiny-GGUF/commit/4d07e0bfc17ea9370951584a54589a7452b2693f","source":"backfill"},{"id":"OBLITERATUS/Qwen3.8-27B-OBLITERATED","date":"2026-08-20T23:50:51.000Z","classes":["template"],"sentence":"Qwen3.8-27B-OBLITERATED: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2026-08-20).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"c3cf9e34abf4f9e3","after":"1bffd744ab18e116"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"c3b1b243d49db803d826fdf0d0d35a3f73340a29","toSha":"fc09b0df0896106d9bd6bcd7403905dd83fe12d7","commit":"https://huggingface.co/OBLITERATUS/Qwen3.8-27B-OBLITERATED/commit/fc09b0df0896106d9bd6bcd7403905dd83fe12d7","source":"backfill"},{"id":"OBLITERATUS/Qwen3.8-27B-OBLITERATED","date":"2026-08-19T17:43:02.000Z","classes":["structural","tokenizer","template"],"sentence":"Qwen3.8-27B-OBLITERATED: derived parameters 26.79B to 26.79B (+7.0M, +0.01 GB of bf16 weights); architectures [\"Qwen3_5ForCausalLM\"] to [\"Qwen3_5ForConditionalGeneration\"]; a different class is instantiated from the same id; a chat template was added; apply_chat_template now formats prompts that it previously refused or formatted by default (2026-08-19).","keys":[{"file":"config","key":"architectures","before":["Qwen3_5ForCausalLM"],"after":["Qwen3_5ForConditionalGeneration"]},{"file":"config","key":"model_type","before":"qwen3_5_text","after":"qwen3_5"},{"file":"tokenizer","key":"add_bos_token","before":null,"after":false},{"file":"tokenizer","key":"chat_template_sha256","before":null,"after":"c3cf9e34abf4f9e3"}],"consequence":{"params":{"before":26786168832,"after":26793185024},"weightsGbBf16":{"before":53.572,"after":53.586},"trainPeakGb":{"before":1557.235,"after":1657.854},"fingerprint":{"before":"40a1bb72","after":"023cc914"},"context":null,"eos":null,"bos":null},"fromSha":"74e7b99a96d1621cb5ffdb6496649e8587fdd64f","toSha":"c3b1b243d49db803d826fdf0d0d35a3f73340a29","commit":"https://huggingface.co/OBLITERATUS/Qwen3.8-27B-OBLITERATED/commit/c3b1b243d49db803d826fdf0d0d35a3f73340a29","source":"backfill"},{"id":"ornith-ai/Ornith-1.5-397B-FP8","date":"2026-08-19T10:29:12.000Z","classes":["structural"],"sentence":"Ornith-1.5-397B-FP8: quantization_config changed; the stored format of the weights is different (2026-08-19).","keys":[{"file":"config","key":"quantization_config","before":{"format":"float-quantized","global_compression_ratio":null,"kv_cache_scheme":null,"quant_method":"compressed-tensors","quantization_status":"compressed","version":"0.17.1","_rest":"29c7aa6d82e6624f"},"after":{"format":"float-quantized","global_compression_ratio":null,"kv_cache_scheme":null,"quant_method":"compressed-tensors","quantization_status":"compressed","version":"0.17.1","_rest":"f12e3515168a7269"}}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"d522d30813cf4e37f5a8b0247888b2bf0d24671c","toSha":"845ced766890c425afb8ea3c57558e370ccdd122","commit":"https://huggingface.co/ornith-ai/Ornith-1.5-397B-FP8/commit/845ced766890c425afb8ea3c57558e370ccdd122","source":"backfill"},{"id":"ornith-ai/Ornith-1.5-35B-A3B-FP8","date":"2026-08-19T09:31:44.000Z","classes":["structural"],"sentence":"Ornith-1.5-35B-A3B-FP8: quantization_config changed; the stored format of the weights is different (2026-08-19).","keys":[{"file":"config","key":"quantization_config","before":{"format":"float-quantized","global_compression_ratio":null,"kv_cache_scheme":null,"quant_method":"compressed-tensors","quantization_status":"compressed","version":"0.17.1","_rest":"2d0a0b633bb45320"},"after":{"format":"float-quantized","global_compression_ratio":null,"kv_cache_scheme":null,"quant_method":"compressed-tensors","quantization_status":"compressed","version":"0.17.1","_rest":"f12e3515168a7269"}}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"87c04e2ed6faffc981020a713405d6fec7f5d4cb","toSha":"a2f6be72cf31b13d85bd69c9ce8ed380d4b82e2f","commit":"https://huggingface.co/ornith-ai/Ornith-1.5-35B-A3B-FP8/commit/a2f6be72cf31b13d85bd69c9ce8ed380d4b82e2f","source":"backfill"},{"id":"ornith-ai/Ornith-1.5-35B-A3B","date":"2026-08-19T02:48:00.000Z","classes":["tokenizer"],"sentence":"Ornith-1.5-35B-A3B: add_bos_token unset to false; encoding the same text yields different ids (2026-08-19).","keys":[{"file":"tokenizer","key":"add_bos_token","before":null,"after":false}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"a111f308b7a9fc99010afd1d412ae92969f854ad","toSha":"97d67d3cbb1ebd7388cea849f81c257f4e2e99a0","commit":"https://huggingface.co/ornith-ai/Ornith-1.5-35B-A3B/commit/97d67d3cbb1ebd7388cea849f81c257f4e2e99a0","source":"backfill"},{"id":"webAI-Official/TwIL-LM3","date":"2026-08-13T18:52:38.000Z","classes":["tokenizer"],"sentence":"TwIL-LM3: tokenizer model_max_length 131072 to 65536; truncation defaults change; tokenizer_class unset to \"PreTrainedTokenizerFast\" (2026-08-13).","keys":[{"file":"tokenizer","key":"model_max_length","before":131072,"after":65536},{"file":"tokenizer","key":"tokenizer_class","before":null,"after":"PreTrainedTokenizerFast"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"e8b809cc93b444a39b5c621c4ed89178b76193f0","toSha":"715bde26ca65dc386eb7a035635182d8f361c096","commit":"https://huggingface.co/webAI-Official/TwIL-LM3/commit/715bde26ca65dc386eb7a035635182d8f361c096","source":"backfill"},{"id":"nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-NVFP4","date":"2026-08-10T18:43:22.000Z","classes":["structural"],"sentence":"NVIDIA-Nemotron-3.5-Lightning-30B-A3B-NVFP4: context window 262144 to 1048576 tokens (2026-08-10).","keys":[{"file":"config","key":"max_position_embeddings","before":262144,"after":1048576}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":{"before":262144,"after":1048576},"eos":null,"bos":null},"fromSha":"3ff6a82aadec49f1f12d8bdb12df5575560d1a44","toSha":"b3d65e6d74e25168350ac9e78805e8fd8e14c0cb","commit":"https://huggingface.co/nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-NVFP4/commit/b3d65e6d74e25168350ac9e78805e8fd8e14c0cb","source":"backfill"},{"id":"nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-NVFP4","date":"2026-08-05T22:25:13.000Z","classes":["template"],"sentence":"NVIDIA-Nemotron-3.5-Lightning-30B-A3B-NVFP4: derived parameters 32.91B to 105.10B (+72.19B, +144.37 GB of bf16 weights); the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2026-08-05).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"82753bef5cedc493","after":"58933db77d3099b4"}],"consequence":{"params":{"before":32913396928,"after":105098907008},"weightsGbBf16":{"before":65.827,"after":210.198},"trainPeakGb":{"before":888.313,"after":2565.477},"fingerprint":{"before":"8bce5cb8","after":"dbf803de"},"context":null,"eos":null,"bos":null},"fromSha":"4a21ff66eed9affcad84f3bd40931a1b41ff90d3","toSha":"3ff6a82aadec49f1f12d8bdb12df5575560d1a44","commit":"https://huggingface.co/nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-NVFP4/commit/3ff6a82aadec49f1f12d8bdb12df5575560d1a44","source":"backfill"},{"id":"nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-BF16","date":"2026-08-05T02:37:28.000Z","classes":["template"],"sentence":"NVIDIA-Nemotron-3.5-Lightning-30B-A3B-BF16: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2026-08-05).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"82753bef5cedc493","after":"58933db77d3099b4"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"b3caaabed0263651a17dc1f2d4ce97e794f76c44","toSha":"742ebb650c8bcf75d8ef92a3b19faabf7dd32664","commit":"https://huggingface.co/nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-BF16/commit/742ebb650c8bcf75d8ef92a3b19faabf7dd32664","source":"backfill"},{"id":"poolside/Laguna-S-2.1-NVFP4","date":"2026-08-01T09:01:58.000Z","classes":["structural"],"sentence":"Laguna-S-2.1-NVFP4: context window 262144 to 1048576 tokens; quantization_config changed; the stored format of the weights is different (2026-08-01).","keys":[{"file":"config","key":"max_position_embeddings","before":262144,"after":1048576},{"file":"config","key":"quantization_config","before":{"format":"nvfp4-pack-quantized","global_compression_ratio":null,"quant_method":"compressed-tensors","quantization_status":"compressed","version":"0.14.1.dev11+gf2ee47b","_rest":"0763e92087092809"},"after":{"format":"nvfp4-pack-quantized","global_compression_ratio":null,"quant_method":"compressed-tensors","quantization_status":"compressed","version":"0.14.1.dev11+gf2ee47b","_rest":"12102894df38f3ac"}}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":{"before":262144,"after":1048576},"eos":null,"bos":null},"fromSha":"07614121b31898586430f189d27a25a0be310843","toSha":"f8fdfcdc4e7b0c474a0102430a8cae0a3a358669","commit":"https://huggingface.co/poolside/Laguna-S-2.1-NVFP4/commit/f8fdfcdc4e7b0c474a0102430a8cae0a3a358669","source":"backfill"},{"id":"RadixArk/Kimi-K3-DSpark","date":"2026-07-31T23:43:06.000Z","classes":["structural"],"sentence":"Kimi-K3-DSpark: rope_scaling added ({\"factor\":16,\"original_max_position_embeddings\":65536,\"ro...); positions are now scaled (2026-07-31).","keys":[{"file":"config","key":"rope_scaling","before":null,"after":{"factor":16,"original_max_position_embeddings":65536,"rope_type":"yarn"}}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"049288a5e88dc5fa18efe9dd00c9f51ba9331fe2","toSha":"9c4b2577dacb572ce88e8aad4357dffb4f6c9796","commit":"https://huggingface.co/RadixArk/Kimi-K3-DSpark/commit/9c4b2577dacb572ce88e8aad4357dffb4f6c9796","source":"backfill"},{"id":"poolside/Laguna-XS-2.1","date":"2026-07-29T13:24:04.000Z","classes":["generation-defaults"],"sentence":"Laguna-XS-2.1: generation defaults changed (top_k unset to 20); a call that does not pass these explicitly now generates differently (2026-07-29).","keys":[{"file":"generation","key":"top_k","before":null,"after":20}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"205dc65dd4bda946c50da6b7522b215734fa107b","toSha":"e9df9a59996d790b94b70f3fef343fe1d9e34bdf","commit":"https://huggingface.co/poolside/Laguna-XS-2.1/commit/e9df9a59996d790b94b70f3fef343fe1d9e34bdf","source":"backfill"},{"id":"k-chirkunov/gemma4-e4b-claims-comparison","date":"2026-07-24T14:43:57.000Z","classes":["tokenizer","template"],"sentence":"gemma4-e4b-claims-comparison: pad token \"<eos>\" to \"<pad>\"; batched training masks a different id; the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2026-07-24).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"2f1b4d75d067bae3","after":"0a2c8073c878ab1d"},{"file":"tokenizer","key":"pad_token","before":"<eos>","after":"<pad>"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"0a74f6a696f42ef40ca7ec44b6e36393fae06f91","toSha":"6c780642f024d4a10192af5a1e1190da514529eb","commit":"https://huggingface.co/k-chirkunov/gemma4-e4b-claims-comparison/commit/6c780642f024d4a10192af5a1e1190da514529eb","source":"backfill"},{"id":"ibm-research/MoLFormer-XL-both-10pct","date":"2026-07-23T14:28:52.000Z","classes":["tokenizer"],"sentence":"MoLFormer-XL-both-10pct: EOS declared as \"<eos>\" (was unset); generation that relied on the library default now stops on this token; BOS declared as \"<bos>\" (was unset); tokenizer model_max_length 1e+30 to unset; truncation defaults change (2026-07-23).","keys":[{"file":"tokenizer","key":"bos_token","before":null,"after":"<bos>"},{"file":"tokenizer","key":"eos_token","before":null,"after":"<eos>"},{"file":"tokenizer","key":"model_max_length","before":1e+30,"after":null},{"file":"tokenizer","key":"tokenizer_class","before":"MolformerTokenizer","after":"PreTrainedTokenizerFast"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":{"before":null,"after":"<eos>"},"bos":{"before":null,"after":"<bos>"}},"fromSha":"abb341ed7cee8c7783088d70bf9f86dff266844e","toSha":"361063d0ad524ef77cf39b08469f6be770dc550f","commit":"https://huggingface.co/ibm-research/MoLFormer-XL-both-10pct/commit/361063d0ad524ef77cf39b08469f6be770dc550f","source":"backfill"},{"id":"poolside/Laguna-S-2.1-NVFP4","date":"2026-07-22T22:20:50.000Z","classes":["structural"],"sentence":"Laguna-S-2.1-NVFP4: quantization_config changed; the stored format of the weights is different (2026-07-22).","keys":[{"file":"config","key":"quantization_config","before":{"format":"nvfp4-pack-quantized","global_compression_ratio":null,"quant_method":"compressed-tensors","quantization_status":"compressed","version":"0.14.1.dev11+gf2ee47b","_rest":"4edf07466fabd016"},"after":{"format":"nvfp4-pack-quantized","global_compression_ratio":null,"quant_method":"compressed-tensors","quantization_status":"compressed","version":"0.14.1.dev11+gf2ee47b","_rest":"0763e92087092809"}}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"b482b5d57fda6e4e562a652869bde24ba2a57c92","toSha":"07614121b31898586430f189d27a25a0be310843","commit":"https://huggingface.co/poolside/Laguna-S-2.1-NVFP4/commit/07614121b31898586430f189d27a25a0be310843","source":"backfill"},{"id":"poolside/Laguna-S-2.1-NVFP4","date":"2026-07-21T22:27:24.000Z","classes":["template","generation-defaults"],"sentence":"Laguna-S-2.1-NVFP4: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering; generation defaults changed (max_new_tokens 32768 to unset); a call that does not pass these explicitly now generates differently (2026-07-21).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"2d3c724b3c2e9eb7","after":"444819b8ad461287"},{"file":"generation","key":"max_new_tokens","before":32768,"after":null}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"ea64155234b727936788ba6be1fd01e5f84e357e","toSha":"b482b5d57fda6e4e562a652869bde24ba2a57c92","commit":"https://huggingface.co/poolside/Laguna-S-2.1-NVFP4/commit/b482b5d57fda6e4e562a652869bde24ba2a57c92","source":"backfill"},{"id":"poolside/Laguna-S-2.1","date":"2026-07-21T22:26:31.000Z","classes":["template","generation-defaults"],"sentence":"Laguna-S-2.1: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering; generation defaults changed (max_new_tokens 32768 to unset); a call that does not pass these explicitly now generates differently (2026-07-21).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"cba88d199f034794","after":"444819b8ad461287"},{"file":"generation","key":"max_new_tokens","before":32768,"after":null}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"6548c108001d16faa203e5a48f52134b64c15709","toSha":"179ee67cf0fff5391c67fe1a392ea849fa6d643f","commit":"https://huggingface.co/poolside/Laguna-S-2.1/commit/179ee67cf0fff5391c67fe1a392ea849fa6d643f","source":"backfill"},{"id":"poolside/Laguna-S-2.1","date":"2026-07-21T21:48:08.000Z","classes":["template"],"sentence":"Laguna-S-2.1: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2026-07-21).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"2d3c724b3c2e9eb7","after":"cba88d199f034794"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"e4daa58f75dbe26f54af634b276e69336beec443","toSha":"fac825ee40ea453f0620403615e94b0f8b6f5c87","commit":"https://huggingface.co/poolside/Laguna-S-2.1/commit/fac825ee40ea453f0620403615e94b0f8b6f5c87","source":"backfill"},{"id":"tencent/Hy3","date":"2026-07-20T07:01:39.000Z","classes":["metadata"],"sentence":"Hy3: declared dtype unset to \"bfloat16\" (2026-07-20).","keys":[{"file":"config","key":"dtype","before":null,"after":"bfloat16"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"3917583e8e630bd692dba80241ca349b8975aac4","toSha":"a960ebc3da325ba167f069f76c41eb62c9280d22","commit":"https://huggingface.co/tencent/Hy3/commit/a960ebc3da325ba167f069f76c41eb62c9280d22","source":"backfill"},{"id":"microsoft/phi-4","date":"2026-07-14T14:22:25.000Z","classes":["tokenizer"],"sentence":"phi-4: EOS changed from 100265 to [100257,100265]; anything fine-tuned or served against the old value stops at a different token (2026-07-14).","keys":[{"file":"generation","key":"eos_token_id","before":100265,"after":[100257,100265]}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":{"before":100265,"after":[100257,100265]},"bos":null},"fromSha":"187ef0342fff0eb3333be9f00389385e95ef0b61","toSha":"2db69c1c3e91a05d2c64a3185acfbaf36f744e25","commit":"https://huggingface.co/microsoft/phi-4/commit/2db69c1c3e91a05d2c64a3185acfbaf36f744e25","source":"backfill"},{"id":"ibm-granite/granite-vision-4.1-4b","date":"2026-07-13T15:35:23.000Z","classes":["structural"],"sentence":"granite-vision-4.1-4b: vocab_size changed from 100353 to 100352; the embedding and output head change shape, so adapters and merges made against the old vocabulary no longer load cleanly; derived parameters 3.32B to 3.32B (-2.6K, -0.00 GB of bf16 weights) (2026-07-13).","keys":[{"file":"config","key":"vocab_size","before":100353,"after":100352}],"consequence":{"params":{"before":3319790080,"after":3319787520},"weightsGbBf16":{"before":6.64,"after":6.64},"trainPeakGb":{"before":263.386,"after":263.386},"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"6eada830c190d274873a8d7ab65a5914fe154b20","toSha":"37d591f06319e8f1638b5adcf58bdf50e0f84f7a","commit":"https://huggingface.co/ibm-granite/granite-vision-4.1-4b/commit/37d591f06319e8f1638b5adcf58bdf50e0f84f7a","source":"backfill"},{"id":"ibm-granite/granite-vision-4.1-4b","date":"2026-07-12T09:28:13.000Z","classes":["tokenizer"],"sentence":"granite-vision-4.1-4b: tokenizer_class \"GPT2Tokenizer\" to \"PreTrainedTokenizerFast\" (2026-07-12).","keys":[{"file":"tokenizer","key":"tokenizer_class","before":"GPT2Tokenizer","after":"PreTrainedTokenizerFast"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"b593b0ae76f89d2ad7b6bc2cbce4ee1de2ccfdd8","toSha":"3710103319078a6709949489c726cad0e5db03be","commit":"https://huggingface.co/ibm-granite/granite-vision-4.1-4b/commit/3710103319078a6709949489c726cad0e5db03be","source":"backfill"},{"id":"OpenMOSS-Team/MOSS-Transcribe-Diarize","date":"2026-07-08T05:25:04.000Z","classes":["tokenizer"],"sentence":"MOSS-Transcribe-Diarize: add_bos_token false to unset; encoding the same text yields different ids (2026-07-08).","keys":[{"file":"tokenizer","key":"add_bos_token","before":false,"after":null}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"6fd616513b960e4aa32d2b08e7d0c4f34849fe40","toSha":"6ed90e43711363f5bb5a9569e459b36bd32a5309","commit":"https://huggingface.co/OpenMOSS-Team/MOSS-Transcribe-Diarize/commit/6ed90e43711363f5bb5a9569e459b36bd32a5309","source":"backfill"},{"id":"poolside/Laguna-XS-2.1","date":"2026-07-07T08:43:44.000Z","classes":["template"],"sentence":"Laguna-XS-2.1: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2026-07-07).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"5f1bf14725b6a0e6","after":"2d3c724b3c2e9eb7"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"c0a69d7d6cc11e551559a841fc20bf9998760e94","toSha":"575f0f288a8e61da7492135a678c52dc371a7936","commit":"https://huggingface.co/poolside/Laguna-XS-2.1/commit/575f0f288a8e61da7492135a678c52dc371a7936","source":"backfill"},{"id":"OpenMOSS-Team/MOSS-Transcribe-Diarize","date":"2026-07-06T16:08:13.000Z","classes":["structural","metadata"],"sentence":"MOSS-Transcribe-Diarize: context window 40960 to 131072 tokens; declared dtype unset to \"bfloat16\" (2026-07-06).","keys":[{"file":"config","key":"dtype","before":null,"after":"bfloat16"},{"file":"config","key":"max_position_embeddings","before":40960,"after":131072}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":{"before":40960,"after":131072},"eos":null,"bos":null},"fromSha":"0844c4adb24300bc7c6cd91e379bc790f939f2d6","toSha":"6fd616513b960e4aa32d2b08e7d0c4f34849fe40","commit":"https://huggingface.co/OpenMOSS-Team/MOSS-Transcribe-Diarize/commit/6fd616513b960e4aa32d2b08e7d0c4f34849fe40","source":"backfill"},{"id":"tencent/Hy3","date":"2026-07-05T05:58:15.000Z","classes":["template"],"sentence":"Hy3: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2026-07-05).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"476bd0982264098c","after":"7fc351fee674c137"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"c1131ba7fe64e8cf0c16fa762a9b8585066caf44","toSha":"3917583e8e630bd692dba80241ca349b8975aac4","commit":"https://huggingface.co/tencent/Hy3/commit/3917583e8e630bd692dba80241ca349b8975aac4","source":"backfill"},{"id":"Bahushruth/Qwen3.6-35B-A3B-abliterated-v4","date":"2026-07-03T09:54:42.000Z","classes":["structural"],"sentence":"Qwen3.6-35B-A3B-abliterated-v4: architectures [\"Qwen3_5MoeForConditionalGeneration\"] to [\"Qwen3_5MoeForCausalLM\"]; a different class is instantiated from the same id (2026-07-03).","keys":[{"file":"config","key":"architectures","before":["Qwen3_5MoeForConditionalGeneration"],"after":["Qwen3_5MoeForCausalLM"]},{"file":"config","key":"model_type","before":"qwen3_5_moe","after":"qwen3_5_moe_text"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"1424051343e80f7fa5a73ebaab9dbcbe159f68cb","toSha":"1c28f5014d7e9c20170af40e5aa028679fb30bfb","commit":"https://huggingface.co/Bahushruth/Qwen3.6-35B-A3B-abliterated-v4/commit/1c28f5014d7e9c20170af40e5aa028679fb30bfb","source":"backfill"},{"id":"Bahushruth/Qwen3.6-35B-A3B-abliterated-v4","date":"2026-07-03T09:11:42.000Z","classes":["structural"],"sentence":"Qwen3.6-35B-A3B-abliterated-v4: architectures [\"Qwen3_5MoeForCausalLM\"] to [\"Qwen3_5MoeForConditionalGeneration\"]; a different class is instantiated from the same id (2026-07-03).","keys":[{"file":"config","key":"architectures","before":["Qwen3_5MoeForCausalLM"],"after":["Qwen3_5MoeForConditionalGeneration"]},{"file":"config","key":"model_type","before":"qwen3_5_moe_text","after":"qwen3_5_moe"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"8f5458e5254ae55f5747d3d70d523cde323c9f75","toSha":"1424051343e80f7fa5a73ebaab9dbcbe159f68cb","commit":"https://huggingface.co/Bahushruth/Qwen3.6-35B-A3B-abliterated-v4/commit/1424051343e80f7fa5a73ebaab9dbcbe159f68cb","source":"backfill"},{"id":"nvidia/Qwen3.6-27B-NVFP4","date":"2026-06-30T14:40:07.000Z","classes":["structural"],"sentence":"Qwen3.6-27B-NVFP4: quantization_config changed; the stored format of the weights is different (2026-06-30).","keys":[{"file":"config","key":"quantization_config","before":{"quant_algo":"MIXED_PRECISION","quant_method":"modelopt","_rest":"7fed6b7d2333e79c"},"after":{"quant_algo":"MIXED_PRECISION","quant_method":"modelopt","_rest":"4ca9a8df034a1c97"}}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"7af26f72b568c66084c0b093b4cbc1efd1c18845","toSha":"0893e1606ff3d5f97a441f405d5fc541a6bdf404","commit":"https://huggingface.co/nvidia/Qwen3.6-27B-NVFP4/commit/0893e1606ff3d5f97a441f405d5fc541a6bdf404","source":"backfill"},{"id":"google/tipsv2-so400m14","date":"2026-06-27T04:30:37.000Z","classes":["structural","tokenizer"],"sentence":"tipsv2-so400m14: hidden size unset to 1152, MLP width unset to 4304, attention heads unset to 16, layer count unset to 27; context window unset to 64 tokens; pad token unset to 0; batched training masks a different id (2026-06-27).","keys":[{"file":"config","key":"hidden_size","before":null,"after":1152},{"file":"config","key":"intermediate_size","before":null,"after":4304},{"file":"config","key":"max_position_embeddings","before":null,"after":64},{"file":"config","key":"num_attention_heads","before":null,"after":16},{"file":"config","key":"num_hidden_layers","before":null,"after":27},{"file":"config","key":"pad_token_id","before":null,"after":0}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":{"before":null,"after":64},"eos":null,"bos":null},"fromSha":"36046283529aa8ccf97cf01587c9f938257756c0","toSha":"e448549aec1f5ade08d4846904e5d489cc181494","commit":"https://huggingface.co/google/tipsv2-so400m14/commit/e448549aec1f5ade08d4846904e5d489cc181494","source":"backfill"},{"id":"google/tipsv2-b14","date":"2026-06-27T04:30:36.000Z","classes":["structural","tokenizer"],"sentence":"tipsv2-b14: hidden size unset to 768, MLP width unset to 3072, attention heads unset to 12, layer count unset to 12; context window unset to 64 tokens; pad token unset to 0; batched training masks a different id (2026-06-27).","keys":[{"file":"config","key":"hidden_size","before":null,"after":768},{"file":"config","key":"intermediate_size","before":null,"after":3072},{"file":"config","key":"max_position_embeddings","before":null,"after":64},{"file":"config","key":"num_attention_heads","before":null,"after":12},{"file":"config","key":"num_hidden_layers","before":null,"after":12},{"file":"config","key":"pad_token_id","before":null,"after":0}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":{"before":null,"after":64},"eos":null,"bos":null},"fromSha":"105fb6406a851d4c6ef851b60fb08e4fd917fed9","toSha":"e79d05f9946e069fb9418cd0c6f3f3c69b5f8110","commit":"https://huggingface.co/google/tipsv2-b14/commit/e79d05f9946e069fb9418cd0c6f3f3c69b5f8110","source":"backfill"},{"id":"ornith-ai/Ornith-1.0-397B-FP8","date":"2026-06-25T13:29:57.000Z","classes":["template"],"sentence":"Ornith-1.0-397B-FP8: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2026-06-25).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"8b4d21a1e70ccbc8","after":"a4aee8afcf2e0711"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"ff7a5d5de72297c8f2b5111ebcb4deecff6d88d4","toSha":"6ae590cf5f2d72ed5d6f34c56798a2ad533d8ff3","commit":"https://huggingface.co/ornith-ai/Ornith-1.0-397B-FP8/commit/6ae590cf5f2d72ed5d6f34c56798a2ad533d8ff3","source":"backfill"},{"id":"nvidia/MiniMax-M3-NVFP4","date":"2026-06-24T08:34:43.000Z","classes":["tokenizer"],"sentence":"MiniMax-M3-NVFP4: EOS declared as \"[e~[\" (was unset); generation that relied on the library default now stops on this token; BOS declared as \"]~b]\" (was unset); pad token unset to \"]!p~[\"; batched training masks a different id (2026-06-24).","keys":[{"file":"tokenizer","key":"bos_token","before":null,"after":"]~b]"},{"file":"tokenizer","key":"eos_token","before":null,"after":"[e~["},{"file":"tokenizer","key":"model_max_length","before":null,"after":40960000},{"file":"tokenizer","key":"pad_token","before":null,"after":"]!p~["},{"file":"tokenizer","key":"tokenizer_class","before":null,"after":"PreTrainedTokenizerFast"},{"file":"tokenizer","key":"unk_token","before":null,"after":"[e~["}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":{"before":null,"after":"[e~["},"bos":{"before":null,"after":"]~b]"}},"fromSha":"98a8de483b8f7c48834132d08c9ed1c9c67216e4","toSha":"3aadd016535c57786220a5c0b11f93ab85bc946f","commit":"https://huggingface.co/nvidia/MiniMax-M3-NVFP4/commit/3aadd016535c57786220a5c0b11f93ab85bc946f","source":"backfill"},{"id":"z-lab/Qwen3.6-35B-A3B-DFlash","date":"2026-06-19T20:54:45.000Z","classes":["structural","tokenizer"],"sentence":"Qwen3.6-35B-A3B-DFlash: EOS changed from 248046 to 248044; anything fine-tuned or served against the old value stops at a different token; layer count 8 to 6, KV heads 4 to 8; derived parameters 1.47B to 1.37B (-100.7M, -0.20 GB of bf16 weights) (2026-06-19).","keys":[{"file":"config","key":"eos_token_id","before":248046,"after":248044},{"file":"config","key":"num_hidden_layers","before":8,"after":6},{"file":"config","key":"num_key_value_heads","before":4,"after":8},{"file":"config","key":"pad_token_id","before":248044,"after":null},{"file":"config","key":"rope_scaling","before":{"beta_fast":32,"beta_slow":1,"factor":64,"original_max_position_embeddings":4096,"rope_type":"yarn","type":"yarn"},"after":null},{"file":"config","key":"sliding_window","before":null,"after":4096}],"consequence":{"params":{"before":1470386688,"after":1369715200},"weightsGbBf16":{"before":2.941,"after":2.739},"trainPeakGb":{"before":167.781,"after":155.82},"fingerprint":{"before":"267aeeb5","after":"fee4da1b"},"context":null,"eos":{"before":248046,"after":248044},"bos":null},"fromSha":"31977fbe13a86e8b961774f773058175676d89b8","toSha":"0ce151d016ba8f0f0454d160a81b52ad4588c924","commit":"https://huggingface.co/z-lab/Qwen3.6-35B-A3B-DFlash/commit/0ce151d016ba8f0f0454d160a81b52ad4588c924","source":"backfill"},{"id":"moonshotai/Kimi-K2.7-Code","date":"2026-06-15T07:49:29.000Z","classes":["template","generation-defaults"],"sentence":"Kimi-K2.7-Code: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering; generation defaults changed (temperature unset to 1, top_p unset to 0.95); a call that does not pass these explicitly now generates differently (2026-06-15).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"1324ac27170d9de6","after":"8d21200dff7e85be"},{"file":"generation","key":"temperature","before":null,"after":1},{"file":"generation","key":"top_p","before":null,"after":0.95}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"da26fe4d239c5c9d14507f1f909894201f911ede","toSha":"74797c9c62378b951a1f6fcf5c4631024e9b8bef","commit":"https://huggingface.co/moonshotai/Kimi-K2.7-Code/commit/74797c9c62378b951a1f6fcf5c4631024e9b8bef","source":"backfill"},{"id":"poolside/Laguna-XS.2","date":"2026-05-27T09:00:30.000Z","classes":["generation-defaults"],"sentence":"Laguna-XS.2: generation defaults changed (max_new_tokens 2048 to 4096, min_p unset to 0, temperature 0.7 to 1, top_p 0.9 to 1); a call that does not pass these explicitly now generates differently (2026-05-27).","keys":[{"file":"generation","key":"max_new_tokens","before":2048,"after":4096},{"file":"generation","key":"min_p","before":null,"after":0},{"file":"generation","key":"temperature","before":0.7,"after":1},{"file":"generation","key":"top_p","before":0.9,"after":1}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"339151dd189fb95fdb2a8d91bd699940c9e0fe10","toSha":"07c6c8e9fc4a43381946f981850a7822c12ad8c5","commit":"https://huggingface.co/poolside/Laguna-XS.2/commit/07c6c8e9fc4a43381946f981850a7822c12ad8c5","source":"backfill"},{"id":"poolside/Laguna-XS.2","date":"2026-05-26T14:59:09.000Z","classes":["structural"],"sentence":"Laguna-XS.2: context window 131072 to 262144 tokens (2026-05-26).","keys":[{"file":"config","key":"max_position_embeddings","before":131072,"after":262144}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":{"before":131072,"after":262144},"eos":null,"bos":null},"fromSha":"7b8e693607ac6653fdcc32a14d53b199332dd9d7","toSha":"339151dd189fb95fdb2a8d91bd699940c9e0fe10","commit":"https://huggingface.co/poolside/Laguna-XS.2/commit/339151dd189fb95fdb2a8d91bd699940c9e0fe10","source":"backfill"},{"id":"moonshotai/Kimi-K2.6","date":"2026-05-12T03:12:21.000Z","classes":["tokenizer"],"sentence":"Kimi-K2.6: tokenizer_class \"TikTokenTokenizerFast\" to \"TikTokenTokenizer\" (2026-05-12).","keys":[{"file":"tokenizer","key":"tokenizer_class","before":"TikTokenTokenizerFast","after":"TikTokenTokenizer"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"81bcaaa7947338ce2641983d98947bca0cc1a4d4","toSha":"b5aabbfb20227ed42becbf5541dbffd213942c58","commit":"https://huggingface.co/moonshotai/Kimi-K2.6/commit/b5aabbfb20227ed42becbf5541dbffd213942c58","source":"backfill"},{"id":"moonshotai/Kimi-K2.6","date":"2026-05-11T09:48:31.000Z","classes":["tokenizer"],"sentence":"Kimi-K2.6: tokenizer_class \"TikTokenTokenizer\" to \"TikTokenTokenizerFast\" (2026-05-11).","keys":[{"file":"tokenizer","key":"tokenizer_class","before":"TikTokenTokenizer","after":"TikTokenTokenizerFast"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"2755962d07cb42aa2d988a35bcb65cd4a9c2de82","toSha":"81bcaaa7947338ce2641983d98947bca0cc1a4d4","commit":"https://huggingface.co/moonshotai/Kimi-K2.6/commit/81bcaaa7947338ce2641983d98947bca0cc1a4d4","source":"backfill"},{"id":"XiaomiMiMo/MiMo-V2.5","date":"2026-05-08T10:22:13.000Z","classes":["template"],"sentence":"MiMo-V2.5: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2026-05-08).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"e2da147933d7d30e","after":"3134ac101acd29d3"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"13b5e3f92ab9572523fa21c7f1bfe9c92228aaca","toSha":"2fd4f899a491de2fb0beeafe32b5d700b251f593","commit":"https://huggingface.co/XiaomiMiMo/MiMo-V2.5/commit/2fd4f899a491de2fb0beeafe32b5d700b251f593","source":"backfill"},{"id":"XiaomiMiMo/MiMo-V2.5","date":"2026-05-07T04:23:16.000Z","classes":["tokenizer"],"sentence":"MiMo-V2.5: EOS changed from [151643,151645,1561672] to [151643,151645,151672]; anything fine-tuned or served against the old value stops at a different token (2026-05-07).","keys":[{"file":"generation","key":"eos_token_id","before":[151643,151645,1561672],"after":[151643,151645,151672]}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":{"before":[151643,151645,1561672],"after":[151643,151645,151672]},"bos":null},"fromSha":"4da2748fc0a0123d817280b1b0cd3682a487a161","toSha":"13b5e3f92ab9572523fa21c7f1bfe9c92228aaca","commit":"https://huggingface.co/XiaomiMiMo/MiMo-V2.5/commit/13b5e3f92ab9572523fa21c7f1bfe9c92228aaca","source":"backfill"},{"id":"nvidia/Gemma-4-26B-A4B-NVFP4","date":"2026-05-07T00:43:59.000Z","classes":["template"],"sentence":"Gemma-4-26B-A4B-NVFP4: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2026-05-07).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"85a08664d16d8f3b","after":"94899c0f917d93f6"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"b0b91da4467f3038cd01c6448bf497c22d951287","toSha":"bbf84e07ec612504a2ad296b3c7ab2e1712f7ea2","commit":"https://huggingface.co/nvidia/Gemma-4-26B-A4B-NVFP4/commit/bbf84e07ec612504a2ad296b3c7ab2e1712f7ea2","source":"backfill"},{"id":"zeroentropy/zerank-2-reranker","date":"2026-05-05T17:30:44.000Z","classes":["template"],"sentence":"zerank-2-reranker: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2026-05-05).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"a55ee1b1660128b7","after":"77137ede3bb2f14f"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"b269504291ee2e4c6ee13bde0d1f422be1c9b0a0","toSha":"d3fb3320b99762de2c520d9c74449cd26e0925e3","commit":"https://huggingface.co/zeroentropy/zerank-2-reranker/commit/d3fb3320b99762de2c520d9c74449cd26e0925e3","source":"backfill"},{"id":"nvidia/Gemma-4-31B-IT-NVFP4","date":"2026-05-01T08:45:55.000Z","classes":["template"],"sentence":"Gemma-4-31B-IT-NVFP4: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2026-05-01).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"85a08664d16d8f3b","after":"94899c0f917d93f6"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"61521ee452a45ae05ca99b3b19fb44df64d36824","toSha":"05fa17010ac2bf68365b33bdd20f07faa10654b4","commit":"https://huggingface.co/nvidia/Gemma-4-31B-IT-NVFP4/commit/05fa17010ac2bf68365b33bdd20f07faa10654b4","source":"backfill"},{"id":"moonshotai/Kimi-K2.5","date":"2026-04-30T03:56:40.000Z","classes":["structural"],"sentence":"Kimi-K2.5: quantization_config changed; the stored format of the weights is different (2026-04-30).","keys":[{"file":"config","key":"quantization_config","before":{"format":"pack-quantized","kv_cache_scheme":null,"quant_method":"compressed-tensors","quantization_status":"compressed","_rest":"f35791990945cd67"},"after":{"format":"pack-quantized","kv_cache_scheme":null,"quant_method":"compressed-tensors","quantization_status":"compressed","_rest":"f6dd5147b0795d0c"}}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"6650964d566fe57e540934e575005af04e933bbe","toSha":"4d01dfe0332d63057c186e0b262165819efb6611","commit":"https://huggingface.co/moonshotai/Kimi-K2.5/commit/4d01dfe0332d63057c186e0b262165819efb6611","source":"backfill"},{"id":"moonshotai/Kimi-K2.6","date":"2026-04-30T03:21:16.000Z","classes":["structural"],"sentence":"Kimi-K2.6: quantization_config changed; the stored format of the weights is different (2026-04-30).","keys":[{"file":"config","key":"quantization_config","before":{"format":"pack-quantized","kv_cache_scheme":null,"quant_method":"compressed-tensors","quantization_status":"compressed","_rest":"f35791990945cd67"},"after":{"format":"pack-quantized","kv_cache_scheme":null,"quant_method":"compressed-tensors","quantization_status":"compressed","_rest":"f6dd5147b0795d0c"}}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"e0ee936b6220f8a475e3d00f03ea7fcdde58c360","toSha":"2755962d07cb42aa2d988a35bcb65cd4a9c2de82","commit":"https://huggingface.co/moonshotai/Kimi-K2.6/commit/2755962d07cb42aa2d988a35bcb65cd4a9c2de82","source":"backfill"},{"id":"sakamakismile/Huihui-Qwen3.6-27B-abliterated-NVFP4-MTP","date":"2026-04-29T00:23:09.000Z","classes":["template"],"sentence":"Huihui-Qwen3.6-27B-abliterated-NVFP4-MTP: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2026-04-29).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"e84f32a23fdda276","after":"c1c0c1df2ff80407"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"637d6567085aedb60d6342fdb9a3420aebe32055","toSha":"9297556d20ede43fcccfef8769a8ebb20e017534","commit":"https://huggingface.co/sakamakismile/Huihui-Qwen3.6-27B-abliterated-NVFP4-MTP/commit/9297556d20ede43fcccfef8769a8ebb20e017534","source":"backfill"},{"id":"ibm-granite/granite-vision-4.1-4b","date":"2026-04-28T19:45:49.000Z","classes":["template"],"sentence":"granite-vision-4.1-4b: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2026-04-28).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"29825b7fcc5ce44c","after":"7047d55e44c123cb"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"3556c946df5cc7458965d50deb4df6ff4e4c3ffe","toSha":"b593b0ae76f89d2ad7b6bc2cbce4ee1de2ccfdd8","commit":"https://huggingface.co/ibm-granite/granite-vision-4.1-4b/commit/b593b0ae76f89d2ad7b6bc2cbce4ee1de2ccfdd8","source":"backfill"},{"id":"XiaomiMiMo/MiMo-V2.5","date":"2026-04-28T12:42:00.000Z","classes":["structural","template"],"sentence":"MiMo-V2.5: context window 262144 to 1048576 tokens; rope_theta 5000000 to 10000000; every position embedding changes; the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2026-04-28).","keys":[{"file":"config","key":"max_position_embeddings","before":262144,"after":1048576},{"file":"config","key":"rope_theta","before":5000000,"after":10000000},{"file":"tokenizer","key":"chat_template_sha256","before":"a342c8de81a9881e","after":"e2da147933d7d30e"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":{"before":262144,"after":1048576},"eos":null,"bos":null},"fromSha":"c3f3abb0657e83f4894e1645f51072f975e62b93","toSha":"4da2748fc0a0123d817280b1b0cd3682a487a161","commit":"https://huggingface.co/XiaomiMiMo/MiMo-V2.5/commit/4da2748fc0a0123d817280b1b0cd3682a487a161","source":"backfill"},{"id":"nvidia/Nemotron-3-Nano-Omni-30B-A3B-Reasoning-BF16","date":"2026-04-28T01:32:23.000Z","classes":["tokenizer"],"sentence":"Nemotron-3-Nano-Omni-30B-A3B-Reasoning-BF16: pad token unset to \"<|im_end|>\"; batched training masks a different id (2026-04-28).","keys":[{"file":"tokenizer","key":"pad_token","before":null,"after":"<|im_end|>"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"23d21acd455d9836d50c48570a329bde77e08ba4","toSha":"9de8c684c189a58757559f852ac3314c609b2080","commit":"https://huggingface.co/nvidia/Nemotron-3-Nano-Omni-30B-A3B-Reasoning-BF16/commit/9de8c684c189a58757559f852ac3314c609b2080","source":"backfill"},{"id":"XiaomiMiMo/MiMo-V2.5","date":"2026-04-27T17:34:17.000Z","classes":["tokenizer","template","generation-defaults"],"sentence":"MiMo-V2.5: EOS changed from [151643,151645] to [151643,151645,1561672]; anything fine-tuned or served against the old value stops at a different token; the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering; generation defaults changed (temperature unset to 1, top_p unset to 0.95); a call that does not pass these explicitly now generates differently (2026-04-27).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"beebeefcded62677","after":"a342c8de81a9881e"},{"file":"generation","key":"eos_token_id","before":[151643,151645],"after":[151643,151645,1561672]},{"file":"generation","key":"temperature","before":null,"after":1},{"file":"generation","key":"top_p","before":null,"after":0.95}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":{"before":[151643,151645],"after":[151643,151645,1561672]},"bos":null},"fromSha":"4f78a4d18f532779ecc6cf3d33ce398a20f78328","toSha":"c3f3abb0657e83f4894e1645f51072f975e62b93","commit":"https://huggingface.co/XiaomiMiMo/MiMo-V2.5/commit/c3f3abb0657e83f4894e1645f51072f975e62b93","source":"backfill"},{"id":"nvidia/Nemotron-3-Nano-Omni-30B-A3B-Reasoning-BF16","date":"2026-04-21T15:38:18.000Z","classes":["template"],"sentence":"Nemotron-3-Nano-Omni-30B-A3B-Reasoning-BF16: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2026-04-21).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"5bf09630a205d95b","after":"41428e0c65e312c3"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"eb24a90335903c7b16951954976ef53715d1fc79","toSha":"23d21acd455d9836d50c48570a329bde77e08ba4","commit":"https://huggingface.co/nvidia/Nemotron-3-Nano-Omni-30B-A3B-Reasoning-BF16/commit/23d21acd455d9836d50c48570a329bde77e08ba4","source":"backfill"},{"id":"nvidia/Nemotron-3-Nano-Omni-30B-A3B-Reasoning-BF16","date":"2026-04-20T15:51:03.000Z","classes":["generation-defaults"],"sentence":"Nemotron-3-Nano-Omni-30B-A3B-Reasoning-BF16: generation defaults changed (do_sample unset to true, max_new_tokens unset to 16384, repetition_penalty unset to 1, temperature unset to 0.6); a call that does not pass these explicitly now generates differently (2026-04-20).","keys":[{"file":"generation","key":"do_sample","before":null,"after":true},{"file":"generation","key":"max_new_tokens","before":null,"after":16384},{"file":"generation","key":"repetition_penalty","before":null,"after":1},{"file":"generation","key":"temperature","before":null,"after":0.6},{"file":"generation","key":"top_p","before":null,"after":0.95}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"d0e46b0fc1ecef1231852dc17039519375fea71d","toSha":"43f3e65ffd905780acc8dc7e277efa3d99da3747","commit":"https://huggingface.co/nvidia/Nemotron-3-Nano-Omni-30B-A3B-Reasoning-BF16/commit/43f3e65ffd905780acc8dc7e277efa3d99da3747","source":"backfill"},{"id":"nvidia/Nemotron-3-Nano-Omni-30B-A3B-Reasoning-BF16","date":"2026-04-20T06:14:15.000Z","classes":["structural"],"sentence":"Nemotron-3-Nano-Omni-30B-A3B-Reasoning-BF16: architectures [\"Nemotron_3_Nano_Omni_30B_A3B_Reasoning\"] to [\"NemotronH_Nano_Omni_Reasoning_V3\"]; a different class is instantiated from the same id (2026-04-20).","keys":[{"file":"config","key":"architectures","before":["Nemotron_3_Nano_Omni_30B_A3B_Reasoning"],"after":["NemotronH_Nano_Omni_Reasoning_V3"]},{"file":"config","key":"model_type","before":"Nemotron_3_Nano_Omni_30B_A3B_Reasoning","after":"NemotronH_Nano_Omni_Reasoning_V3"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"62f3c9dfb13f23ac930f100c72d360ca331d2aa5","toSha":"d0e46b0fc1ecef1231852dc17039519375fea71d","commit":"https://huggingface.co/nvidia/Nemotron-3-Nano-Omni-30B-A3B-Reasoning-BF16/commit/d0e46b0fc1ecef1231852dc17039519375fea71d","source":"backfill"},{"id":"MiniMaxAI/MiniMax-M2.7","date":"2026-04-20T04:28:14.000Z","classes":["structural"],"sentence":"MiniMax-M2.7: context window 196608 to 204800 tokens (2026-04-20).","keys":[{"file":"config","key":"max_position_embeddings","before":196608,"after":204800}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":{"before":196608,"after":204800},"eos":null,"bos":null},"fromSha":"d7153aa1bfdcad43f7aa6f7fc8ddeb2e8da49306","toSha":"d494266a4affc0d2995ba1fa35c8481cbd84294b","commit":"https://huggingface.co/MiniMaxAI/MiniMax-M2.7/commit/d494266a4affc0d2995ba1fa35c8481cbd84294b","source":"backfill"},{"id":"nvidia/Gemma-4-31B-IT-NVFP4","date":"2026-04-18T08:11:47.000Z","classes":["template"],"sentence":"Gemma-4-31B-IT-NVFP4: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2026-04-18).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"2dfbfc7d538912f4","after":"85a08664d16d8f3b"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"50687fbbadad4eff4f88f6ca3cec7bf99af871f5","toSha":"61521ee452a45ae05ca99b3b19fb44df64d36824","commit":"https://huggingface.co/nvidia/Gemma-4-31B-IT-NVFP4/commit/61521ee452a45ae05ca99b3b19fb44df64d36824","source":"backfill"},{"id":"z-lab/Qwen3.6-35B-A3B-DFlash","date":"2026-04-17T02:22:30.000Z","classes":["structural"],"sentence":"Qwen3.6-35B-A3B-DFlash: rope_scaling rewritten in a different format ({\"beta_fast\":32,\"beta_slow\":1,\"factor\":64,\"original_max_p... to {\"beta_fast\":32,\"beta_slow\":1,\"factor\":64,\"original_max_p...); a loader that only reads the old format rejects or ignores the new one (2026-04-17).","keys":[{"file":"config","key":"rope_scaling","before":{"beta_fast":32,"beta_slow":1,"factor":64,"original_max_position_embeddings":4096},"after":{"beta_fast":32,"beta_slow":1,"factor":64,"original_max_position_embeddings":4096,"rope_type":"yarn","type":"yarn"}}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"1f71bf5b8b3d13c6068284ca02b2c046cc0b5aed","toSha":"92297e6cfd224a858c3ba24281f7c43ae6c1bb4a","commit":"https://huggingface.co/z-lab/Qwen3.6-35B-A3B-DFlash/commit/92297e6cfd224a858c3ba24281f7c43ae6c1bb4a","source":"backfill"},{"id":"zai-org/GLM-5.1","date":"2026-04-16T06:11:43.000Z","classes":["template"],"sentence":"GLM-5.1: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2026-04-16).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"df60c73492112e64","after":"03b1bbff20331e54"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"f5ed80ea7943de93bdb66223a637d7c3af6d8fe4","toSha":"442b5279043b5f43f382003e7272731f680e1393","commit":"https://huggingface.co/zai-org/GLM-5.1/commit/442b5279043b5f43f382003e7272731f680e1393","source":"backfill"},{"id":"nvidia/llama-nemotron-rerank-1b-v2","date":"2026-04-10T11:03:00.000Z","classes":["structural"],"sentence":"llama-nemotron-rerank-1b-v2: rope_scaling rewritten in a different format ({\"factor\":32,\"high_freq_factor\":4,\"low_freq_factor\":1,\"or... to {\"factor\":32,\"high_freq_factor\":4,\"low_freq_factor\":1,\"or...); a loader that only reads the old format rejects or ignores the new one (2026-04-10).","keys":[{"file":"config","key":"rope_scaling","before":{"factor":32,"high_freq_factor":4,"low_freq_factor":1,"original_max_position_embeddings":8192,"rope_type":"llama3"},"after":{"factor":32,"high_freq_factor":4,"low_freq_factor":1,"original_max_position_embeddings":8192,"rope_type":"llama3","rope_theta":500000}}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"c0bcb4c755dd28faaf0738fe0dd7ca7c9a251ac5","toSha":"8fd3e5d962d44cfe65d4ba0784eebed44cf136b0","commit":"https://huggingface.co/nvidia/llama-nemotron-rerank-1b-v2/commit/8fd3e5d962d44cfe65d4ba0784eebed44cf136b0","source":"backfill"},{"id":"MiniMaxAI/MiniMax-M2.7","date":"2026-04-09T14:17:34.000Z","classes":["template"],"sentence":"MiniMax-M2.7: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2026-04-09).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"6197838b2be77135","after":"893d908f7b5cc65f"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"4c9387fe99ab54b9478ab9ee3c1d5f19623bded5","toSha":"d7153aa1bfdcad43f7aa6f7fc8ddeb2e8da49306","commit":"https://huggingface.co/MiniMaxAI/MiniMax-M2.7/commit/d7153aa1bfdcad43f7aa6f7fc8ddeb2e8da49306","source":"backfill"},{"id":"MiniMaxAI/MiniMax-M2.7","date":"2026-04-09T14:11:22.000Z","classes":["metadata"],"sentence":"MiniMax-M2.7: declared dtype unset to \"bfloat16\" (2026-04-09).","keys":[{"file":"config","key":"dtype","before":null,"after":"bfloat16"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"a0ca8c65a93219ea256d47fd0851c49d13fca5fc","toSha":"4c9387fe99ab54b9478ab9ee3c1d5f19623bded5","commit":"https://huggingface.co/MiniMaxAI/MiniMax-M2.7/commit/4c9387fe99ab54b9478ab9ee3c1d5f19623bded5","source":"backfill"},{"id":"jinaai/jina-reranker-m0","date":"2026-04-09T07:14:14.000Z","classes":["template"],"sentence":"jina-reranker-m0: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2026-04-09).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"a0bc6f6fc7a29a80","after":"4ac7f9da4bf4242a"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"a93850af27e4a262b254461c27c9336850ca07c4","toSha":"94bfe0aeb2d4dd7978362699cddd5893d4e0adc8","commit":"https://huggingface.co/jinaai/jina-reranker-m0/commit/94bfe0aeb2d4dd7978362699cddd5893d4e0adc8","source":"backfill"},{"id":"nomic-ai/nomic-embed-text-v1.5","date":"2026-04-07T14:17:02.000Z","classes":["structural","tokenizer"],"sentence":"nomic-embed-text-v1.5: head dim unset to 64; derived parameters 108.9M to 110.1M (+1.2M, +0.00 GB of bf16 weights); context window 8192 to 2048 tokens (2026-04-07).","keys":[{"file":"config","key":"head_dim","before":null,"after":64},{"file":"config","key":"max_position_embeddings","before":8192,"after":2048},{"file":"config","key":"pad_token_id","before":null,"after":0},{"file":"config","key":"rope_theta","before":null,"after":1000}],"consequence":{"params":{"before":108874752,"after":110054400},"weightsGbBf16":{"before":0.218,"after":0.22},"trainPeakGb":{"before":7.3,"after":22.951},"fingerprint":null,"context":{"before":8192,"after":2048},"eos":null,"bos":null},"fromSha":"e5cf08aadaa33385f5990def41f7a23405aec398","toSha":"e9b6763023c676ca8431644204f50c2b100d9aab","commit":"https://huggingface.co/nomic-ai/nomic-embed-text-v1.5/commit/e9b6763023c676ca8431644204f50c2b100d9aab","source":"backfill"},{"id":"nomic-ai/nomic-embed-text-v1","date":"2026-04-07T14:16:41.000Z","classes":["structural","tokenizer"],"sentence":"nomic-embed-text-v1: head dim unset to 64; derived parameters 108.9M to 114.8M (+5.9M, +0.01 GB of bf16 weights); rope_scaling added ({\"factor\":2,\"rope_type\":\"dynamic\"}); positions are now scaled (2026-04-07).","keys":[{"file":"config","key":"head_dim","before":null,"after":64},{"file":"config","key":"pad_token_id","before":null,"after":0},{"file":"config","key":"rope_scaling","before":null,"after":{"rope_type":"dynamic","factor":2}},{"file":"config","key":"rope_theta","before":null,"after":1000}],"consequence":{"params":{"before":108874752,"after":114772992},"weightsGbBf16":{"before":0.218,"after":0.23},"trainPeakGb":{"before":7.3,"after":23.042},"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"cc62377b015c53a3bf52bb2f4eb8c55326d3f162","toSha":"3ac47f125a41961d13b397d0332866be2f9152e1","commit":"https://huggingface.co/nomic-ai/nomic-embed-text-v1/commit/3ac47f125a41961d13b397d0332866be2f9152e1","source":"backfill"},{"id":"nvidia/Gemma-4-31B-IT-NVFP4","date":"2026-04-02T16:57:10.000Z","classes":["tokenizer"],"sentence":"Gemma-4-31B-IT-NVFP4: pad token \"<eos>\" to \"<pad>\"; batched training masks a different id (2026-04-02).","keys":[{"file":"tokenizer","key":"pad_token","before":"<eos>","after":"<pad>"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"5d24e54bc06049c87faa380094a4959d2951d3c9","toSha":"4caa7d91810d0856081b8660df86d5baaa78ec70","commit":"https://huggingface.co/nvidia/Gemma-4-31B-IT-NVFP4/commit/4caa7d91810d0856081b8660df86d5baaa78ec70","source":"backfill"},{"id":"nvidia/llama-nemotron-embed-1b-v2","date":"2026-03-30T13:20:41.000Z","classes":["structural"],"sentence":"llama-nemotron-embed-1b-v2: rope_scaling rewritten in a different format ({\"factor\":32,\"high_freq_factor\":4,\"low_freq_factor\":1,\"or... to {\"factor\":32,\"high_freq_factor\":4,\"low_freq_factor\":1,\"or...); a loader that only reads the old format rejects or ignores the new one (2026-03-30).","keys":[{"file":"config","key":"rope_scaling","before":{"factor":32,"high_freq_factor":4,"low_freq_factor":1,"original_max_position_embeddings":8192,"rope_type":"llama3"},"after":{"factor":32,"high_freq_factor":4,"low_freq_factor":1,"original_max_position_embeddings":8192,"rope_type":"llama3","rope_theta":500000}}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"46b6d1019f1857f3a9bbf07b90f7372dc111c8d2","toSha":"b4caa8456edd360b3b4e938d94ed4398dd437fad","commit":"https://huggingface.co/nvidia/llama-nemotron-embed-1b-v2/commit/b4caa8456edd360b3b4e938d94ed4398dd437fad","source":"backfill"},{"id":"nvidia/NVIDIA-Nemotron-3-Nano-4B-BF16","date":"2026-03-16T16:41:07.000Z","classes":["generation-defaults"],"sentence":"NVIDIA-Nemotron-3-Nano-4B-BF16: generation defaults changed (temperature unset to 1, top_p unset to 0.95); a call that does not pass these explicitly now generates differently (2026-03-16).","keys":[{"file":"generation","key":"temperature","before":null,"after":1},{"file":"generation","key":"top_p","before":null,"after":0.95}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"63811f6dfe321499ec633ac3c4d394b25de18162","toSha":"b1ac4a3f7049cc760ad2e893d95f63174b4af89e","commit":"https://huggingface.co/nvidia/NVIDIA-Nemotron-3-Nano-4B-BF16/commit/b1ac4a3f7049cc760ad2e893d95f63174b4af89e","source":"backfill"},{"id":"zai-org/GLM-5","date":"2026-03-11T12:23:19.000Z","classes":["template"],"sentence":"GLM-5: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2026-03-11).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"d63ad536c3c81880","after":"45a1d5635afd3d8e"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"22ce68ad9a4707fad5356209fee6c54ae2c4cd87","toSha":"f4c624070fb778e07ad16fb04c34dad055be3fce","commit":"https://huggingface.co/zai-org/GLM-5/commit/f4c624070fb778e07ad16fb04c34dad055be3fce","source":"backfill"},{"id":"stepfun-ai/Step-3.5-Flash","date":"2026-03-09T03:54:54.000Z","classes":["structural"],"sentence":"Step-3.5-Flash: derived parameters 8.28B to 8.81B (+528.1M, +1.06 GB of bf16 weights); tie_word_embeddings unset to false; the output head is no longer shared with the embedding (2026-03-09).","keys":[{"file":"config","key":"tie_word_embeddings","before":null,"after":false}],"consequence":{"params":{"before":8281427968,"after":8809523072},"weightsGbBf16":{"before":16.563,"after":17.619},"trainPeakGb":{"before":565.738,"after":623.181},"fingerprint":{"before":"2de5a957","after":"1773acdc"},"context":null,"eos":null,"bos":null},"fromSha":"7eb38d5f01854a401901ea6800bd9328f91a5085","toSha":"ef9cf3cda3b4d76a67250e6eaa11d6ad7a8f2a73","commit":"https://huggingface.co/stepfun-ai/Step-3.5-Flash/commit/ef9cf3cda3b4d76a67250e6eaa11d6ad7a8f2a73","source":"backfill"},{"id":"sarvamai/sarvam-30b","date":"2026-03-06T08:08:31.000Z","classes":["structural"],"sentence":"sarvam-30b: context window 32768 to 131072 tokens (2026-03-06).","keys":[{"file":"config","key":"max_position_embeddings","before":32768,"after":131072}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":{"before":32768,"after":131072},"eos":null,"bos":null},"fromSha":"501c3a733d172c48e2b00f8fdec3f25196fca167","toSha":"19a5eb7aa4d9810ebe8cec497ee0cdb7a3db68f0","commit":"https://huggingface.co/sarvamai/sarvam-30b/commit/19a5eb7aa4d9810ebe8cec497ee0cdb7a3db68f0","source":"backfill"},{"id":"nvidia/Kimi-K2.5-NVFP4","date":"2026-03-04T18:11:13.000Z","classes":["structural"],"sentence":"Kimi-K2.5-NVFP4: quantization_config changed; the stored format of the weights is different (2026-03-04).","keys":[{"file":"config","key":"quantization_config","before":{"quant_algo":"NVFP4","quant_method":"modelopt","_rest":"25548a8a87a6132f"},"after":{"quant_algo":"NVFP4","quant_method":"modelopt","_rest":"df3d08bba5e8236d"}}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"dd47e2388cd35fb37a2ebf03b938b351ae0ffc35","toSha":"e4e908c073784de20ad3af0be653421f1088922d","commit":"https://huggingface.co/nvidia/Kimi-K2.5-NVFP4/commit/e4e908c073784de20ad3af0be653421f1088922d","source":"backfill"},{"id":"sarvamai/sarvam-30b","date":"2026-03-04T16:00:03.000Z","classes":["structural"],"sentence":"sarvam-30b: context window 131072 to 32768 tokens (2026-03-04).","keys":[{"file":"config","key":"max_position_embeddings","before":131072,"after":32768}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":{"before":131072,"after":32768},"eos":null,"bos":null},"fromSha":"1c398d1fb429273ddf457174341a7ab5bcba6a10","toSha":"501c3a733d172c48e2b00f8fdec3f25196fca167","commit":"https://huggingface.co/sarvamai/sarvam-30b/commit/501c3a733d172c48e2b00f8fdec3f25196fca167","source":"backfill"},{"id":"amd/DeepSeek-R1-MXFP4","date":"2026-02-27T01:36:16.000Z","classes":["structural"],"sentence":"DeepSeek-R1-MXFP4: quantization_config changed; the stored format of the weights is different (2026-02-27).","keys":[{"file":"config","key":"quantization_config","before":{"algo_config":null,"quant_method":"quark","quant_mode":"eager_mode","softmax_quant_spec":null,"_rest":"995d9e5e61000aff"},"after":{"algo_config":null,"quant_method":"quark","quant_mode":"eager_mode","softmax_quant_spec":null,"_rest":"67e4766e2d5c3cfa"}}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"359d1ee20ef45aa480c19dec855f534cf622a03e","toSha":"d31023d90e8da554e9ca98e3b0cbea85be1652e3","commit":"https://huggingface.co/amd/DeepSeek-R1-MXFP4/commit/d31023d90e8da554e9ca98e3b0cbea85be1652e3","source":"backfill"},{"id":"jinaai/jina-embeddings-v5-text-nano-retrieval","date":"2026-02-19T15:45:24.000Z","classes":["tokenizer"],"sentence":"jina-embeddings-v5-text-nano-retrieval: tokenizer_class \"TokenizersBackend\" to \"PreTrainedTokenizer\" (2026-02-19).","keys":[{"file":"tokenizer","key":"tokenizer_class","before":"TokenizersBackend","after":"PreTrainedTokenizer"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"fde746ac2c676a1df07ca4688f9e0c294606a0f1","toSha":"51c1540f0289b0fc88e5858abcd92039a09250aa","commit":"https://huggingface.co/jinaai/jina-embeddings-v5-text-nano-retrieval/commit/51c1540f0289b0fc88e5858abcd92039a09250aa","source":"backfill"},{"id":"jinaai/jina-embeddings-v5-text-nano-retrieval","date":"2026-02-16T11:16:58.000Z","classes":["structural","tokenizer"],"sentence":"jina-embeddings-v5-text-nano-retrieval: rope_scaling removed (was {\"factor\":1,\"rope_type\":\"linear\"}); positions are no longer scaled; tokenizer_class \"PreTrainedTokenizerFast\" to \"TokenizersBackend\" (2026-02-16).","keys":[{"file":"config","key":"rope_scaling","before":{"factor":1,"rope_type":"linear"},"after":null},{"file":"tokenizer","key":"tokenizer_class","before":"PreTrainedTokenizerFast","after":"TokenizersBackend"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"e15a71f0ccefcdeb9beb01b301ea7f828eb26850","toSha":"fde746ac2c676a1df07ca4688f9e0c294606a0f1","commit":"https://huggingface.co/jinaai/jina-embeddings-v5-text-nano-retrieval/commit/fde746ac2c676a1df07ca4688f9e0c294606a0f1","source":"backfill"},{"id":"jinaai/jina-embeddings-v5-text-nano-retrieval","date":"2026-02-16T10:31:45.000Z","classes":["structural"],"sentence":"jina-embeddings-v5-text-nano-retrieval: model_type \"jina_embeddings_v5\" to \"eurobert\"; a different class is instantiated from the same id (2026-02-16).","keys":[{"file":"config","key":"model_type","before":"jina_embeddings_v5","after":"eurobert"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"849566083b114d8c1722173bb412ee6381908493","toSha":"e15a71f0ccefcdeb9beb01b301ea7f828eb26850","commit":"https://huggingface.co/jinaai/jina-embeddings-v5-text-nano-retrieval/commit/e15a71f0ccefcdeb9beb01b301ea7f828eb26850","source":"backfill"},{"id":"MiniMaxAI/MiniMax-M2.5","date":"2026-02-12T16:07:45.000Z","classes":["template"],"sentence":"MiniMax-M2.5: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2026-02-12).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"6197838b2be77135","after":"f83cec52bb3a7e5a"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"8c436e563d876e41e089f8d24aa481e2f2978be4","toSha":"134ae7c46b8c0152819a9c534dad8628066b23b1","commit":"https://huggingface.co/MiniMaxAI/MiniMax-M2.5/commit/134ae7c46b8c0152819a9c534dad8628066b23b1","source":"backfill"},{"id":"jinaai/jina-embeddings-v5-text-nano-retrieval","date":"2026-02-11T10:35:23.000Z","classes":["structural"],"sentence":"jina-embeddings-v5-text-nano-retrieval: architectures [\"EuroBertForMaskedLM\"] to [\"EuroBertModel\"]; a different class is instantiated from the same id (2026-02-11).","keys":[{"file":"config","key":"architectures","before":["EuroBertForMaskedLM"],"after":["EuroBertModel"]}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"afa39c0f576c3e1465b8ff65aea4a95544de34b0","toSha":"849566083b114d8c1722173bb412ee6381908493","commit":"https://huggingface.co/jinaai/jina-embeddings-v5-text-nano-retrieval/commit/849566083b114d8c1722173bb412ee6381908493","source":"backfill"},{"id":"jinaai/jina-embeddings-v5-text-nano-retrieval","date":"2026-02-10T13:26:25.000Z","classes":["structural"],"sentence":"jina-embeddings-v5-text-nano-retrieval: architectures [\"EuroBertModel\"] to [\"EuroBertForMaskedLM\"]; a different class is instantiated from the same id (2026-02-10).","keys":[{"file":"config","key":"architectures","before":["EuroBertModel"],"after":["EuroBertForMaskedLM"]}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"6f649a5c83c4340e40b0a59e209289237fc1ef26","toSha":"afa39c0f576c3e1465b8ff65aea4a95544de34b0","commit":"https://huggingface.co/jinaai/jina-embeddings-v5-text-nano-retrieval/commit/afa39c0f576c3e1465b8ff65aea4a95544de34b0","source":"backfill"},{"id":"jinaai/jina-embeddings-v5-text-small","date":"2026-02-02T16:00:06.000Z","classes":["structural"],"sentence":"jina-embeddings-v5-text-small: derived parameters 596.0M to 483.0M (-113.0M, -0.23 GB of bf16 weights); context window 40960 to 32768 tokens; model_type \"qwen3\" to \"jina_embeddings_v5\"; a different class is instantiated from the same id (2026-02-02).","keys":[{"file":"config","key":"max_position_embeddings","before":40960,"after":32768},{"file":"config","key":"model_type","before":"qwen3","after":"jina_embeddings_v5"}],"consequence":{"params":{"before":596041728,"after":483024896},"weightsGbBf16":{"before":1.192,"after":0.966},"trainPeakGb":{"before":97.26,"after":63.336},"fingerprint":{"before":"22e368b4","after":"02884d4d"},"context":{"before":40960,"after":32768},"eos":null,"bos":null},"fromSha":"c9756bfc6a062e6d2a993463b28a07aef08ed99d","toSha":"dab81cb52e4164a72665cb7ac7ef5cf98f8e7561","commit":"https://huggingface.co/jinaai/jina-embeddings-v5-text-small/commit/dab81cb52e4164a72665cb7ac7ef5cf98f8e7561","source":"backfill"},{"id":"jinaai/jina-embeddings-v5-text-nano","date":"2026-02-02T16:00:01.000Z","classes":["structural"],"sentence":"jina-embeddings-v5-text-nano: rope_theta 250000 to 1000000; every position embedding changes (2026-02-02).","keys":[{"file":"config","key":"rope_theta","before":250000,"after":1000000}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"b7558cd9bd1f8ace55f7bc9f7daa52c9197f913a","toSha":"371f6f7ce126801c3c8a8cd7d4185938ad557ecf","commit":"https://huggingface.co/jinaai/jina-embeddings-v5-text-nano/commit/371f6f7ce126801c3c8a8cd7d4185938ad557ecf","source":"backfill"},{"id":"moonshotai/Kimi-K2.5","date":"2026-01-29T09:11:59.000Z","classes":["template"],"sentence":"Kimi-K2.5: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2026-01-29).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"e226722cfc8f8dfd","after":"7c91e4b3578b518b"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"70a8ad02d86dbb7459abd22b8a4feca294d2c993","toSha":"6650964d566fe57e540934e575005af04e933bbe","commit":"https://huggingface.co/moonshotai/Kimi-K2.5/commit/6650964d566fe57e540934e575005af04e933bbe","source":"backfill"},{"id":"sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2","date":"2026-01-28T10:02:26.000Z","classes":["tokenizer"],"sentence":"paraphrase-multilingual-MiniLM-L12-v2: tokenizer_class unset to \"PreTrainedTokenizerFast\" (2026-01-28).","keys":[{"file":"tokenizer","key":"tokenizer_class","before":null,"after":"PreTrainedTokenizerFast"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"e62509716f15c5fd03a6fd3156a4bc5e43f83f26","toSha":"e8f8c211226b894fcb81acc59f3b34ba3efd5f42","commit":"https://huggingface.co/sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2/commit/e8f8c211226b894fcb81acc59f3b34ba3efd5f42","source":"backfill"},{"id":"nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-FP8","date":"2025-12-15T09:57:44.000Z","classes":["template"],"sentence":"NVIDIA-Nemotron-3-Nano-30B-A3B-FP8: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2025-12-15).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"b80135b85225bef1","after":"ab7813c3abdd9cb6"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"f9a1d7f631fac131de8fa0bc929545c1eb406135","toSha":"127ab9cc01d7425647f370dca0a5d0ec9dc45e6c","commit":"https://huggingface.co/nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-FP8/commit/127ab9cc01d7425647f370dca0a5d0ec9dc45e6c","source":"backfill"},{"id":"nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-BF16","date":"2025-12-15T09:57:14.000Z","classes":["template"],"sentence":"NVIDIA-Nemotron-3-Nano-30B-A3B-BF16: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2025-12-15).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"b80135b85225bef1","after":"ab7813c3abdd9cb6"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"a7d6c572c46df00ce702a0039b02a0c2572cc61a","toSha":"4fed69b4f37cde78542ef42c69bc567ddca0d645","commit":"https://huggingface.co/nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-BF16/commit/4fed69b4f37cde78542ef42c69bc567ddca0d645","source":"backfill"},{"id":"allenai/Olmo-3-7B-Think","date":"2025-12-10T21:46:11.000Z","classes":["generation-defaults"],"sentence":"Olmo-3-7B-Think: generation defaults changed (do_sample unset to true); a call that does not pass these explicitly now generates differently (2025-12-10).","keys":[{"file":"generation","key":"do_sample","before":null,"after":true}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"032a160e7ffc30f73fad9d7cfcaf56649cbd065c","toSha":"5d89684927443a2bf68dff246f38bbc72458aabe","commit":"https://huggingface.co/allenai/Olmo-3-7B-Think/commit/5d89684927443a2bf68dff246f38bbc72458aabe","source":"backfill"},{"id":"MiniMaxAI/MiniMax-M2","date":"2025-12-08T08:48:48.000Z","classes":["template"],"sentence":"MiniMax-M2: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2025-12-08).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"1deca769f6945a7d","after":"86cd6e574dec52bf"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"9354a5cc95c7206e3371a772b4c8a4f000163401","toSha":"f62f643a718e8be697da52e7d5d8b8d12b991001","commit":"https://huggingface.co/MiniMaxAI/MiniMax-M2/commit/f62f643a718e8be697da52e7d5d8b8d12b991001","source":"backfill"},{"id":"nvidia/Qwen2.5-VL-7B-Instruct-NVFP4","date":"2025-12-06T00:51:49.000Z","classes":["structural"],"sentence":"Qwen2.5-VL-7B-Instruct-NVFP4: quantization_config changed; the stored format of the weights is different (2025-12-06).","keys":[{"file":"config","key":"quantization_config","before":{"quant_algo":"NVFP4","quant_method":"modelopt","_rest":"e751732c9786f88c"},"after":{"quant_algo":"NVFP4","quant_method":"modelopt","_rest":"1c6f232296946273"}}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"7b5d4d62308c86c9feb816b420c15b38f17018f6","toSha":"d13bb1f2d8fbbd9f16cbb7688c8d6b6932d797d7","commit":"https://huggingface.co/nvidia/Qwen2.5-VL-7B-Instruct-NVFP4/commit/d13bb1f2d8fbbd9f16cbb7688c8d6b6932d797d7","source":"backfill"},{"id":"Qwen/Qwen3-Coder-30B-A3B-Instruct-FP8","date":"2025-12-03T08:20:23.000Z","classes":["structural","tokenizer"],"sentence":"Qwen3-Coder-30B-A3B-Instruct-FP8: BOS declared as 151643 (was unset); MLP width 5472 to 6144; quantization_config changed; the stored format of the weights is different (2025-12-03).","keys":[{"file":"config","key":"bos_token_id","before":null,"after":151643},{"file":"config","key":"intermediate_size","before":5472,"after":6144},{"file":"config","key":"quantization_config","before":{"activation_scheme":"dynamic","quant_method":"fp8","_rest":"81cc3297a7cc5d80"},"after":{"activation_scheme":"dynamic","fmt":"e4m3","quant_method":"fp8","_rest":"81cc3297a7cc5d80"}}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":{"before":null,"after":151643}},"fromSha":"e8ab3f2db9e388999a004eea5a31c16a8b517bc0","toSha":"dcaee4d4dfc5ee71ad501f01f530e5652438fde0","commit":"https://huggingface.co/Qwen/Qwen3-Coder-30B-A3B-Instruct-FP8/commit/dcaee4d4dfc5ee71ad501f01f530e5652438fde0","source":"backfill"},{"id":"Qwen/Qwen3-Coder-30B-A3B-Instruct","date":"2025-12-03T08:05:17.000Z","classes":["structural","tokenizer"],"sentence":"Qwen3-Coder-30B-A3B-Instruct: BOS declared as 151643 (was unset); MLP width 5472 to 6144 (2025-12-03).","keys":[{"file":"config","key":"bos_token_id","before":null,"after":151643},{"file":"config","key":"intermediate_size","before":5472,"after":6144}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":{"before":null,"after":151643}},"fromSha":"573fa3901e5799703b1e60825b0ec024a4c0f1d3","toSha":"b2cff646eb4bb1d68355c01b18ae02e7cf42d120","commit":"https://huggingface.co/Qwen/Qwen3-Coder-30B-A3B-Instruct/commit/b2cff646eb4bb1d68355c01b18ae02e7cf42d120","source":"backfill"},{"id":"deepseek-ai/DeepSeek-V3.2","date":"2025-12-01T10:47:01.000Z","classes":["generation-defaults"],"sentence":"DeepSeek-V3.2: generation defaults changed (temperature 0.6 to 1); a call that does not pass these explicitly now generates differently (2025-12-01).","keys":[{"file":"generation","key":"temperature","before":0.6,"after":1}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"fba8af16b2323abda4dffc7508468b44809bacc6","toSha":"c69397ecfd1fd142e90e3fbad51f4c7e40b9f3d3","commit":"https://huggingface.co/deepseek-ai/DeepSeek-V3.2/commit/c69397ecfd1fd142e90e3fbad51f4c7e40b9f3d3","source":"backfill"},{"id":"allenai/Olmo-3-7B-Instruct","date":"2025-11-26T02:06:50.000Z","classes":["generation-defaults"],"sentence":"Olmo-3-7B-Instruct: generation defaults changed (do_sample unset to true); a call that does not pass these explicitly now generates differently (2025-11-26).","keys":[{"file":"generation","key":"do_sample","before":null,"after":true}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"48683dfc961ce0f02f802f109b261d6028d83e0c","toSha":"51c1f0b28546f9cf537583ea5b9f5da4fed23ebe","commit":"https://huggingface.co/allenai/Olmo-3-7B-Instruct/commit/51c1f0b28546f9cf537583ea5b9f5da4fed23ebe","source":"backfill"},{"id":"allenai/Olmo-3-7B-Instruct","date":"2025-11-21T19:13:08.000Z","classes":["generation-defaults"],"sentence":"Olmo-3-7B-Instruct: generation defaults changed (max_new_tokens unset to 32768, temperature unset to 0.6, top_p unset to 0.95); a call that does not pass these explicitly now generates differently (2025-11-21).","keys":[{"file":"generation","key":"max_new_tokens","before":null,"after":32768},{"file":"generation","key":"temperature","before":null,"after":0.6},{"file":"generation","key":"top_p","before":null,"after":0.95}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"d42dc9de51ee481e02b8a35d8222b86c6d7c5b20","toSha":"48683dfc961ce0f02f802f109b261d6028d83e0c","commit":"https://huggingface.co/allenai/Olmo-3-7B-Instruct/commit/48683dfc961ce0f02f802f109b261d6028d83e0c","source":"backfill"},{"id":"allenai/Olmo-3-7B-Think","date":"2025-11-21T19:10:12.000Z","classes":["generation-defaults"],"sentence":"Olmo-3-7B-Think: generation defaults changed (max_new_tokens unset to 32768, temperature unset to 0.6, top_p unset to 0.95); a call that does not pass these explicitly now generates differently (2025-11-21).","keys":[{"file":"generation","key":"max_new_tokens","before":null,"after":32768},{"file":"generation","key":"temperature","before":null,"after":0.6},{"file":"generation","key":"top_p","before":null,"after":0.95}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"923f0a31a295dab8f9c8ee120aab4d7e70e097de","toSha":"032a160e7ffc30f73fad9d7cfcaf56649cbd065c","commit":"https://huggingface.co/allenai/Olmo-3-7B-Think/commit/032a160e7ffc30f73fad9d7cfcaf56649cbd065c","source":"backfill"},{"id":"swiss-ai/Apertus-8B-Instruct-2509","date":"2025-11-14T11:00:08.000Z","classes":["template"],"sentence":"Apertus-8B-Instruct-2509: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2025-11-14).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"00067a2e29980bd1","after":"56f72c27fa2e5653"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"ffb69ffeb76480f975934e7c997feb857151c83b","toSha":"cdb3e4f4ad41e0cc394bb92c302ac2eed57e9586","commit":"https://huggingface.co/swiss-ai/Apertus-8B-Instruct-2509/commit/cdb3e4f4ad41e0cc394bb92c302ac2eed57e9586","source":"backfill"},{"id":"MiniMaxAI/MiniMax-M2","date":"2025-11-07T08:56:23.000Z","classes":["template"],"sentence":"MiniMax-M2: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2025-11-07).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"6197838b2be77135","after":"1deca769f6945a7d"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"78ba357af9df61620ef179ffb2b4c43ea41c2cba","toSha":"9354a5cc95c7206e3371a772b4c8a4f000163401","commit":"https://huggingface.co/MiniMaxAI/MiniMax-M2/commit/9354a5cc95c7206e3371a772b4c8a4f000163401","source":"backfill"},{"id":"MiniMaxAI/MiniMax-M2","date":"2025-11-06T03:30:28.000Z","classes":["structural","tokenizer"],"sentence":"MiniMax-M2: EOS declared as 200020 (was unset); generation that relied on the library default now stops on this token; BOS declared as 200019 (was unset); model_type \"minimax\" to \"minimax_m2\"; a different class is instantiated from the same id (2025-11-06).","keys":[{"file":"config","key":"model_type","before":"minimax","after":"minimax_m2"},{"file":"config","key":"quantization_config","before":{"activation_scheme":"dynamic","fmt":"float8_e4m3fn","quant_method":"fp8","_rest":"0f7293707cf5d1ba"},"after":{"activation_scheme":"dynamic","fmt":"float8_e4m3fn","quant_method":"fp8","_rest":"b46e96439073befc"}},{"file":"generation","key":"bos_token_id","before":null,"after":200019},{"file":"generation","key":"eos_token_id","before":null,"after":200020}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":{"before":null,"after":200020},"bos":{"before":null,"after":200019}},"fromSha":"44cefa66f81ec7ffebcb580c02985726ef4d829e","toSha":"78ba357af9df61620ef179ffb2b4c43ea41c2cba","commit":"https://huggingface.co/MiniMaxAI/MiniMax-M2/commit/78ba357af9df61620ef179ffb2b4c43ea41c2cba","source":"backfill"},{"id":"allenai/Olmo-3-7B-Think","date":"2025-10-31T18:05:35.000Z","classes":["structural"],"sentence":"Olmo-3-7B-Think: architectures [\"Olmo2RetrofitForCausalLM\"] to [\"Olmo3ForCausalLM\"]; a different class is instantiated from the same id (2025-10-31).","keys":[{"file":"config","key":"architectures","before":["Olmo2RetrofitForCausalLM"],"after":["Olmo3ForCausalLM"]},{"file":"config","key":"model_type","before":"olmo2-retrofit","after":"olmo3"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"361b459dd6325b67ceb7f5636db15ba46a93cf62","toSha":"923f0a31a295dab8f9c8ee120aab4d7e70e097de","commit":"https://huggingface.co/allenai/Olmo-3-7B-Think/commit/923f0a31a295dab8f9c8ee120aab4d7e70e097de","source":"backfill"},{"id":"MiniMaxAI/MiniMax-M2","date":"2025-10-27T08:51:03.000Z","classes":["structural"],"sentence":"MiniMax-M2: model_type \"mixtral\" to \"minimax\"; a different class is instantiated from the same id (2025-10-27).","keys":[{"file":"config","key":"model_type","before":"mixtral","after":"minimax"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"e9774173c852cfbca23856fcee6233431d2815f8","toSha":"44cefa66f81ec7ffebcb580c02985726ef4d829e","commit":"https://huggingface.co/MiniMaxAI/MiniMax-M2/commit/44cefa66f81ec7ffebcb580c02985726ef4d829e","source":"backfill"},{"id":"MiniMaxAI/MiniMax-M2","date":"2025-10-27T02:43:12.000Z","classes":["generation-defaults"],"sentence":"MiniMax-M2: generation defaults changed (top_k unset to 40); a call that does not pass these explicitly now generates differently (2025-10-27).","keys":[{"file":"generation","key":"top_k","before":null,"after":40}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"cab79a3e1dc02bab6ffcc80134dee602c2cf583c","toSha":"e9774173c852cfbca23856fcee6233431d2815f8","commit":"https://huggingface.co/MiniMaxAI/MiniMax-M2/commit/e9774173c852cfbca23856fcee6233431d2815f8","source":"backfill"},{"id":"MiniMaxAI/MiniMax-M2","date":"2025-10-25T07:09:48.000Z","classes":["structural"],"sentence":"MiniMax-M2: architectures [\"MiniMaxM2MiniForCausalLM\"] to [\"MiniMaxM2ForCausalLM\"]; a different class is instantiated from the same id (2025-10-25).","keys":[{"file":"config","key":"architectures","before":["MiniMaxM2MiniForCausalLM"],"after":["MiniMaxM2ForCausalLM"]}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"49303277636d105a85ac09729935c61fd0ab17de","toSha":"da5eb8f2eb03c4af9998e0acc47cc21b8ed4c189","commit":"https://huggingface.co/MiniMaxAI/MiniMax-M2/commit/da5eb8f2eb03c4af9998e0acc47cc21b8ed4c189","source":"backfill"},{"id":"moonshotai/Kimi-K2-Instruct","date":"2025-10-22T16:26:41.000Z","classes":["template"],"sentence":"Kimi-K2-Instruct: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2025-10-22).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"966ce6ca6aeee6b8","after":"39e8c195b474c1a4"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"cc613312db6692a12f650552166d3bf1d09e936a","toSha":"0102674b179db4ca5a28cd9a4fb446f87f0c1454","commit":"https://huggingface.co/moonshotai/Kimi-K2-Instruct/commit/0102674b179db4ca5a28cd9a4fb446f87f0c1454","source":"backfill"},{"id":"moonshotai/Kimi-K2-Instruct","date":"2025-10-22T03:27:44.000Z","classes":["template"],"sentence":"Kimi-K2-Instruct: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2025-10-22).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"0cd4cf36e2abadce","after":"966ce6ca6aeee6b8"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"874f2bbf365c6ad034c302fc2e72e617a97e1012","toSha":"cc613312db6692a12f650552166d3bf1d09e936a","commit":"https://huggingface.co/moonshotai/Kimi-K2-Instruct/commit/cc613312db6692a12f650552166d3bf1d09e936a","source":"backfill"},{"id":"moonshotai/Kimi-K2-Instruct","date":"2025-10-10T06:31:51.000Z","classes":["template"],"sentence":"Kimi-K2-Instruct: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2025-10-10).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"0d8c1b9b65099ad3","after":"0cd4cf36e2abadce"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"c52f808f632c07eb8361388616b1d04749373a94","toSha":"874f2bbf365c6ad034c302fc2e72e617a97e1012","commit":"https://huggingface.co/moonshotai/Kimi-K2-Instruct/commit/874f2bbf365c6ad034c302fc2e72e617a97e1012","source":"backfill"},{"id":"nvidia/Qwen2.5-VL-7B-Instruct-NVFP4","date":"2025-10-08T17:43:27.000Z","classes":["structural"],"sentence":"Qwen2.5-VL-7B-Instruct-NVFP4: rope_scaling values changed ({\"mrope_section\":[16,24,24],\"rope_type\":\"default\",\"type\":... to {\"mrope_section\":[16,24,24],\"rope_type\":\"default\",\"type\":...); positions are computed differently (2025-10-08).","keys":[{"file":"config","key":"rope_scaling","before":{"mrope_section":[16,24,24],"rope_type":"default","type":"default"},"after":{"mrope_section":[16,24,24],"rope_type":"default","type":"mrope"}}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"39a5817f4677b8696d6720c9357db8d5b1fd77de","toSha":"7b5d4d62308c86c9feb816b420c15b38f17018f6","commit":"https://huggingface.co/nvidia/Qwen2.5-VL-7B-Instruct-NVFP4/commit/7b5d4d62308c86c9feb816b420c15b38f17018f6","source":"backfill"},{"id":"amd/DeepSeek-R1-MXFP4","date":"2025-10-04T00:44:35.000Z","classes":["structural"],"sentence":"DeepSeek-R1-MXFP4: quantization_config changed; the stored format of the weights is different (2025-10-04).","keys":[{"file":"config","key":"quantization_config","before":{"algo_config":null,"quant_method":"quark","quant_mode":"eager_mode","softmax_quant_spec":null,"_rest":"4294f32e17157b88"},"after":{"algo_config":null,"quant_method":"quark","quant_mode":"eager_mode","softmax_quant_spec":null,"_rest":"995d9e5e61000aff"}}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"a8b84c8ab78875ec0b8a5c15c6bf683b3cc50a53","toSha":"359d1ee20ef45aa480c19dec855f534cf622a03e","commit":"https://huggingface.co/amd/DeepSeek-R1-MXFP4/commit/359d1ee20ef45aa480c19dec855f534cf622a03e","source":"backfill"},{"id":"ai-sage/Giga-Embeddings-instruct","date":"2025-09-23T12:00:06.000Z","classes":["structural","tokenizer"],"sentence":"Giga-Embeddings-instruct: model_type \"gigar\" to \"gigarembed\"; a different class is instantiated from the same id; declared dtype \"bfloat16\" to \"float32\"; tokenizer_class \"PreTrainedTokenizerFast\" to \"PreTrainedTokenizer\" (2025-09-23).","keys":[{"file":"config","key":"dtype","before":"bfloat16","after":"float32"},{"file":"config","key":"model_type","before":"gigar","after":"gigarembed"},{"file":"tokenizer","key":"tokenizer_class","before":"PreTrainedTokenizerFast","after":"PreTrainedTokenizer"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"b3dc287fece8fcb002a9bc02138818b6fdc0b9b8","toSha":"3c12a846f3f964c38cfa04b3f9e95e4f187c97c8","commit":"https://huggingface.co/ai-sage/Giga-Embeddings-instruct/commit/3c12a846f3f964c38cfa04b3f9e95e4f187c97c8","source":"backfill"},{"id":"ai-sage/Giga-Embeddings-instruct","date":"2025-09-23T10:38:32.000Z","classes":["structural","template"],"sentence":"Giga-Embeddings-instruct: head dim 128 to 64, KV heads 2 to 16; derived parameters 3.30B to 2.50B (-801.0M, -1.60 GB of bf16 weights); rope_theta 1300 to 100000; every position embedding changes (2025-09-23).","keys":[{"file":"config","key":"dtype","before":"float32","after":"bfloat16"},{"file":"config","key":"head_dim","before":128,"after":64},{"file":"config","key":"model_type","before":"gigarembed","after":"gigar"},{"file":"config","key":"num_key_value_heads","before":2,"after":16},{"file":"config","key":"rope_theta","before":1300,"after":100000},{"file":"tokenizer","key":"chat_template_sha256","before":null,"after":"ee5407800568eadf"}],"consequence":{"params":{"before":3300146432,"after":2499144704},"weightsGbBf16":{"before":6.6,"after":4.998},"trainPeakGb":{"before":375.903,"after":248.184},"fingerprint":{"before":"87540c74","after":"be837245"},"context":null,"eos":null,"bos":null},"fromSha":"f4891741cf49fc2e724a14188470cd57c39fec93","toSha":"b3dc287fece8fcb002a9bc02138818b6fdc0b9b8","commit":"https://huggingface.co/ai-sage/Giga-Embeddings-instruct/commit/b3dc287fece8fcb002a9bc02138818b6fdc0b9b8","source":"backfill"},{"id":"PrimeIntellect/Qwen3-0.6B","date":"2025-09-22T09:44:23.000Z","classes":["structural","template"],"sentence":"Qwen3-0.6B: declared dtype \"float32\" to \"bfloat16\"; the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2025-09-22).","keys":[{"file":"config","key":"dtype","before":"float32","after":"bfloat16"},{"file":"tokenizer","key":"chat_template_sha256","before":"cd8e9439f0570856","after":"12062a547e3e7fa2"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"1f818745f470b18c51d477e185aa671da451bdfd","toSha":"d26dbbb7fce164b397e53b2dcbc45abf40433c1b","commit":"https://huggingface.co/PrimeIntellect/Qwen3-0.6B/commit/d26dbbb7fce164b397e53b2dcbc45abf40433c1b","source":"backfill"},{"id":"Qwen/Qwen3-4B-Instruct-2507-FP8","date":"2025-09-17T06:57:02.000Z","classes":["tokenizer","template"],"sentence":"Qwen3-4B-Instruct-2507-FP8: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering; tokenizer model_max_length 262144 to 1010000; truncation defaults change (2025-09-17).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"40c21f34cf67d8c7","after":"64f85b198065d0fb"},{"file":"tokenizer","key":"model_max_length","before":262144,"after":1010000}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"6b8fc6f96470133b40f3110e46d25098632c71cd","toSha":"8591804019c8b22094c3b5b4454e0edc05dffc98","commit":"https://huggingface.co/Qwen/Qwen3-4B-Instruct-2507-FP8/commit/8591804019c8b22094c3b5b4454e0edc05dffc98","source":"backfill"},{"id":"Qwen/Qwen3-4B-Instruct-2507","date":"2025-09-17T06:56:53.000Z","classes":["tokenizer","template"],"sentence":"Qwen3-4B-Instruct-2507: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering; tokenizer model_max_length 262144 to 1010000; truncation defaults change (2025-09-17).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"40c21f34cf67d8c7","after":"64f85b198065d0fb"},{"file":"tokenizer","key":"model_max_length","before":262144,"after":1010000}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"1b4199c4f36b0cef378bfb12390c18780c18af4c","toSha":"cdbee75f17c01a7cc42f958dc650907174af0554","commit":"https://huggingface.co/Qwen/Qwen3-4B-Instruct-2507/commit/cdbee75f17c01a7cc42f958dc650907174af0554","source":"backfill"},{"id":"Qwen/Qwen3-30B-A3B-Instruct-2507","date":"2025-09-17T06:56:43.000Z","classes":["template"],"sentence":"Qwen3-30B-A3B-Instruct-2507: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2025-09-17).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"40c21f34cf67d8c7","after":"64f85b198065d0fb"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"3ffd1f50b179e643d839c86df9ffbbefcb0d5018","toSha":"0d7cf23991f47feeb3a57ecb4c9cee8ea4a17bfe","commit":"https://huggingface.co/Qwen/Qwen3-30B-A3B-Instruct-2507/commit/0d7cf23991f47feeb3a57ecb4c9cee8ea4a17bfe","source":"backfill"},{"id":"Qwen/Qwen3-30B-A3B-Instruct-2507-FP8","date":"2025-09-17T06:56:32.000Z","classes":["tokenizer","template"],"sentence":"Qwen3-30B-A3B-Instruct-2507-FP8: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering; tokenizer model_max_length 262144 to 1010000; truncation defaults change (2025-09-17).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"40c21f34cf67d8c7","after":"64f85b198065d0fb"},{"file":"tokenizer","key":"model_max_length","before":262144,"after":1010000}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"81e34e368e9e39d3e6967d27fef73607a3ed96f0","toSha":"5a5a776300a41aaa681dd7ff0106608ef2bc90db","commit":"https://huggingface.co/Qwen/Qwen3-30B-A3B-Instruct-2507-FP8/commit/5a5a776300a41aaa681dd7ff0106608ef2bc90db","source":"backfill"},{"id":"Qwen/Qwen3-235B-A22B-Instruct-2507-FP8","date":"2025-09-17T06:53:28.000Z","classes":["tokenizer","template"],"sentence":"Qwen3-235B-A22B-Instruct-2507-FP8: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering; tokenizer model_max_length 262144 to 1010000; truncation defaults change (2025-09-17).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"40c21f34cf67d8c7","after":"64f85b198065d0fb"},{"file":"tokenizer","key":"model_max_length","before":262144,"after":1010000}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"652b9036a3d93443261dfe5ed47192ba859f14d3","toSha":"e156cb4efae43fbee1a1ab073f946a1377e6b969","commit":"https://huggingface.co/Qwen/Qwen3-235B-A22B-Instruct-2507-FP8/commit/e156cb4efae43fbee1a1ab073f946a1377e6b969","source":"backfill"},{"id":"OpenGVLab/InternVL3-8B","date":"2025-09-11T12:50:07.000Z","classes":["template"],"sentence":"InternVL3-8B: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2025-09-11).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"cabb34545cf3d04f","after":"09a5c48046d889ee"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"9e7a9d574e4861a2a46a642045cbe8f2e514514e","toSha":"853e3a797a661694b1b8ece0cb72dc2b23e3dac9","commit":"https://huggingface.co/OpenGVLab/InternVL3-8B/commit/853e3a797a661694b1b8ece0cb72dc2b23e3dac9","source":"backfill"},{"id":"OpenGVLab/InternVL3-1B","date":"2025-09-11T12:49:40.000Z","classes":["template"],"sentence":"InternVL3-1B: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2025-09-11).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"cabb34545cf3d04f","after":"09a5c48046d889ee"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"aed473f1758b073136496560128af4cfe3984844","toSha":"4415a3b810e636d11dfa86b0e9ba40bb00535aa8","commit":"https://huggingface.co/OpenGVLab/InternVL3-1B/commit/4415a3b810e636d11dfa86b0e9ba40bb00535aa8","source":"backfill"},{"id":"OpenGVLab/InternVL2_5-4B","date":"2025-09-11T12:47:44.000Z","classes":["template"],"sentence":"InternVL2_5-4B: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2025-09-11).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"cabb34545cf3d04f","after":"09a5c48046d889ee"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"de2c1f985d9abc47634b932e19e63b39f70b8da1","toSha":"2cf4a8158bbc40d35015e7c63b527890de4d27b3","commit":"https://huggingface.co/OpenGVLab/InternVL2_5-4B/commit/2cf4a8158bbc40d35015e7c63b527890de4d27b3","source":"backfill"},{"id":"OpenGVLab/InternVL2_5-4B","date":"2025-09-11T10:28:48.000Z","classes":["template"],"sentence":"InternVL2_5-4B: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2025-09-11).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"cd8e9439f0570856","after":"cabb34545cf3d04f"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"f01b74c11c5317424c85e41a9b1d06a940e80c4d","toSha":"de2c1f985d9abc47634b932e19e63b39f70b8da1","commit":"https://huggingface.co/OpenGVLab/InternVL2_5-4B/commit/de2c1f985d9abc47634b932e19e63b39f70b8da1","source":"backfill"},{"id":"OpenGVLab/InternVL3-1B","date":"2025-09-11T10:17:27.000Z","classes":["template"],"sentence":"InternVL3-1B: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2025-09-11).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"cd8e9439f0570856","after":"cabb34545cf3d04f"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"698c83feb064c47aac7a2043d11e4db176c27438","toSha":"aed473f1758b073136496560128af4cfe3984844","commit":"https://huggingface.co/OpenGVLab/InternVL3-1B/commit/aed473f1758b073136496560128af4cfe3984844","source":"backfill"},{"id":"OpenGVLab/InternVL3-8B","date":"2025-09-11T10:10:23.000Z","classes":["template"],"sentence":"InternVL3-8B: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2025-09-11).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"cd8e9439f0570856","after":"cabb34545cf3d04f"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"dab7194eaadae9ff191fef49b961847a18b4c822","toSha":"9e7a9d574e4861a2a46a642045cbe8f2e514514e","commit":"https://huggingface.co/OpenGVLab/InternVL3-8B/commit/9e7a9d574e4861a2a46a642045cbe8f2e514514e","source":"backfill"},{"id":"swiss-ai/Apertus-8B-Instruct-2509","date":"2025-09-01T22:39:33.000Z","classes":["template"],"sentence":"Apertus-8B-Instruct-2509: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2025-09-01).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"9d2ae9dcd0ce7417","after":"00067a2e29980bd1"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"55ec663ac96b82bfa30608b9c7db4926e09591b0","toSha":"ffb69ffeb76480f975934e7c997feb857151c83b","commit":"https://huggingface.co/swiss-ai/Apertus-8B-Instruct-2509/commit/ffb69ffeb76480f975934e7c997feb857151c83b","source":"backfill"},{"id":"swiss-ai/Apertus-8B-Instruct-2509","date":"2025-09-01T15:35:48.000Z","classes":["template"],"sentence":"Apertus-8B-Instruct-2509: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2025-09-01).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"9c269c91d80cf02c","after":"9d2ae9dcd0ce7417"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"fa7d5040df695aa4690450568ef220282cebba33","toSha":"55ec663ac96b82bfa30608b9c7db4926e09591b0","commit":"https://huggingface.co/swiss-ai/Apertus-8B-Instruct-2509/commit/55ec663ac96b82bfa30608b9c7db4926e09591b0","source":"backfill"},{"id":"deepseek-ai/DeepSeek-V3.1","date":"2025-08-26T08:14:11.000Z","classes":["structural"],"sentence":"DeepSeek-V3.1: quantization_config changed; the stored format of the weights is different (2025-08-26).","keys":[{"file":"config","key":"quantization_config","before":{"activation_scheme":"dynamic","fmt":"e4m3","quant_method":"fp8","_rest":"0f7293707cf5d1ba"},"after":{"activation_scheme":"dynamic","fmt":"e4m3","quant_method":"fp8","scale_fmt":"ue8m0","_rest":"0f7293707cf5d1ba"}}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"c7fe4623a4fd23debe09bb40e04e46f88f4b2ddc","toSha":"9e6c48c3fa6bb3e1cf684675dc02e813ca45d20f","commit":"https://huggingface.co/deepseek-ai/DeepSeek-V3.1/commit/9e6c48c3fa6bb3e1cf684675dc02e813ca45d20f","source":"backfill"},{"id":"nvidia/NVIDIA-Nemotron-Nano-9B-v2","date":"2025-08-21T15:42:44.000Z","classes":["tokenizer"],"sentence":"NVIDIA-Nemotron-Nano-9B-v2: tokenizer model_max_length 1e+30 to 131072; truncation defaults change (2025-08-21).","keys":[{"file":"tokenizer","key":"model_max_length","before":1e+30,"after":131072}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"7659b750c2aa7b6ba76937ac4c4c64f4a56fb79f","toSha":"d99d9740c5ac3067aaae19d4771781f667505fda","commit":"https://huggingface.co/nvidia/NVIDIA-Nemotron-Nano-9B-v2/commit/d99d9740c5ac3067aaae19d4771781f667505fda","source":"backfill"},{"id":"Qwen/Qwen3-Coder-480B-A35B-Instruct-FP8","date":"2025-08-21T10:40:35.000Z","classes":["template"],"sentence":"Qwen3-Coder-480B-A35B-Instruct-FP8: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2025-08-21).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"7e0648e7a298d820","after":"5a38bfa058332662"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"1f6f99b0d0f8e209a7574a946b440b1020a6f31b","toSha":"003f183a92fbe5b9a8325aaa8b2ae797c91dd90f","commit":"https://huggingface.co/Qwen/Qwen3-Coder-480B-A35B-Instruct-FP8/commit/003f183a92fbe5b9a8325aaa8b2ae797c91dd90f","source":"backfill"},{"id":"Qwen/Qwen3-Coder-30B-A3B-Instruct-FP8","date":"2025-08-21T10:22:22.000Z","classes":["template"],"sentence":"Qwen3-Coder-30B-A3B-Instruct-FP8: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2025-08-21).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"7e0648e7a298d820","after":"5a38bfa058332662"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"acc0a21828d8d3483ddd1ce54d2e98a89ca80309","toSha":"e8ab3f2db9e388999a004eea5a31c16a8b517bc0","commit":"https://huggingface.co/Qwen/Qwen3-Coder-30B-A3B-Instruct-FP8/commit/e8ab3f2db9e388999a004eea5a31c16a8b517bc0","source":"backfill"},{"id":"Qwen/Qwen3-Coder-30B-A3B-Instruct","date":"2025-08-21T10:18:32.000Z","classes":["template"],"sentence":"Qwen3-Coder-30B-A3B-Instruct: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2025-08-21).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"7e0648e7a298d820","after":"5a38bfa058332662"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"8110d98ad896c995124f3bc44b2f8dd695a4d252","toSha":"573fa3901e5799703b1e60825b0ec024a4c0f1d3","commit":"https://huggingface.co/Qwen/Qwen3-Coder-30B-A3B-Instruct/commit/573fa3901e5799703b1e60825b0ec024a4c0f1d3","source":"backfill"},{"id":"nvidia/NVIDIA-Nemotron-Nano-9B-v2","date":"2025-08-15T04:03:41.000Z","classes":["tokenizer","template"],"sentence":"NVIDIA-Nemotron-Nano-9B-v2: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering; add_bos_token true to false; encoding the same text yields different ids (2025-08-15).","keys":[{"file":"tokenizer","key":"add_bos_token","before":true,"after":false},{"file":"tokenizer","key":"chat_template_sha256","before":"b64f57c601cc2f85","after":"b7a3a520a4bc1bea"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"85507799cec2b7978deff4329546aff29456f916","toSha":"b27e3876f3db535003a7635f4fc20c4ed77e02fa","commit":"https://huggingface.co/nvidia/NVIDIA-Nemotron-Nano-9B-v2/commit/b27e3876f3db535003a7635f4fc20c4ed77e02fa","source":"backfill"},{"id":"openai/gpt-oss-120b","date":"2025-08-13T23:23:29.000Z","classes":["tokenizer"],"sentence":"gpt-oss-120b: EOS changed from [200002,199999] to [200002,199999,200012]; anything fine-tuned or served against the old value stops at a different token (2025-08-13).","keys":[{"file":"generation","key":"eos_token_id","before":[200002,199999],"after":[200002,199999,200012]}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":{"before":[200002,199999],"after":[200002,199999,200012]},"bos":null},"fromSha":"bc75b44b8a2a116a0e4c6659bcd1b7969885f423","toSha":"8b193b0ef83bd41b40eb71fee8f1432315e02a3e","commit":"https://huggingface.co/openai/gpt-oss-120b/commit/8b193b0ef83bd41b40eb71fee8f1432315e02a3e","source":"backfill"},{"id":"openai/gpt-oss-20b","date":"2025-08-13T23:23:06.000Z","classes":["tokenizer"],"sentence":"gpt-oss-20b: EOS changed from [200002,199999] to [200002,199999,200012]; anything fine-tuned or served against the old value stops at a different token (2025-08-13).","keys":[{"file":"generation","key":"eos_token_id","before":[200002,199999],"after":[200002,199999,200012]}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":{"before":[200002,199999],"after":[200002,199999,200012]},"bos":null},"fromSha":"cbf31f62664d4b1360b3a78427f7b3c3ed8f0fa8","toSha":"d666cf3b67006cf8227666739edf25164aaffdeb","commit":"https://huggingface.co/openai/gpt-oss-20b/commit/d666cf3b67006cf8227666739edf25164aaffdeb","source":"backfill"},{"id":"moonshotai/Kimi-K2-Instruct","date":"2025-08-11T13:45:09.000Z","classes":["template"],"sentence":"Kimi-K2-Instruct: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2025-08-11).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"aa8b293348ddc405","after":"0d8c1b9b65099ad3"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"b215bf7476a2287207d8f7f519eaccb136144512","toSha":"c52f808f632c07eb8361388616b1d04749373a94","commit":"https://huggingface.co/moonshotai/Kimi-K2-Instruct/commit/c52f808f632c07eb8361388616b1d04749373a94","source":"backfill"},{"id":"moonshotai/Kimi-K2-Instruct","date":"2025-08-11T05:50:11.000Z","classes":["template"],"sentence":"Kimi-K2-Instruct: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2025-08-11).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"454036503ee2fc14","after":"aa8b293348ddc405"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"6a1f5e6be5b1ddfd220836faed223dd70fbc091e","toSha":"b215bf7476a2287207d8f7f519eaccb136144512","commit":"https://huggingface.co/moonshotai/Kimi-K2-Instruct/commit/b215bf7476a2287207d8f7f519eaccb136144512","source":"backfill"},{"id":"openai/gpt-oss-20b","date":"2025-08-08T22:46:37.000Z","classes":["template"],"sentence":"gpt-oss-20b: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2025-08-08).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"b474759b33c21342","after":"a4c9919cbbd4acdd"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"10e9d713f8e4a9281c59c40be6c58537480635ea","toSha":"cbf31f62664d4b1360b3a78427f7b3c3ed8f0fa8","commit":"https://huggingface.co/openai/gpt-oss-20b/commit/cbf31f62664d4b1360b3a78427f7b3c3ed8f0fa8","source":"backfill"},{"id":"openai/gpt-oss-120b","date":"2025-08-08T22:46:29.000Z","classes":["template"],"sentence":"gpt-oss-120b: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2025-08-08).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"b474759b33c21342","after":"a4c9919cbbd4acdd"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"988e56b6fc3c723ad30aa7f18db742f54f921567","toSha":"bc75b44b8a2a116a0e4c6659bcd1b7969885f423","commit":"https://huggingface.co/openai/gpt-oss-120b/commit/bc75b44b8a2a116a0e4c6659bcd1b7969885f423","source":"backfill"},{"id":"Qwen/Qwen3-30B-A3B-Instruct-2507","date":"2025-08-08T07:20:23.000Z","classes":["tokenizer"],"sentence":"Qwen3-30B-A3B-Instruct-2507: tokenizer model_max_length 262144 to 1010000; truncation defaults change (2025-08-08).","keys":[{"file":"tokenizer","key":"model_max_length","before":262144,"after":1010000}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"97f3f7d6f068582355b6fb44014a1b60e282636b","toSha":"3ffd1f50b179e643d839c86df9ffbbefcb0d5018","commit":"https://huggingface.co/Qwen/Qwen3-30B-A3B-Instruct-2507/commit/3ffd1f50b179e643d839c86df9ffbbefcb0d5018","source":"backfill"},{"id":"openai/gpt-oss-120b","date":"2025-08-07T17:43:56.000Z","classes":["template"],"sentence":"gpt-oss-120b: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2025-08-07).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"f8d9255777615591","after":"b474759b33c21342"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"f39accdb091b6c5710dbbf02674c27064e35019b","toSha":"988e56b6fc3c723ad30aa7f18db742f54f921567","commit":"https://huggingface.co/openai/gpt-oss-120b/commit/988e56b6fc3c723ad30aa7f18db742f54f921567","source":"backfill"},{"id":"openai/gpt-oss-20b","date":"2025-08-07T17:43:45.000Z","classes":["template"],"sentence":"gpt-oss-20b: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2025-08-07).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"f8d9255777615591","after":"b474759b33c21342"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"ec854da5735f125fe36f080d8013482590f9ad7d","toSha":"10e9d713f8e4a9281c59c40be6c58537480635ea","commit":"https://huggingface.co/openai/gpt-oss-20b/commit/10e9d713f8e4a9281c59c40be6c58537480635ea","source":"backfill"},{"id":"Qwen/Qwen3-Coder-30B-A3B-Instruct","date":"2025-08-07T07:08:04.000Z","classes":["template"],"sentence":"Qwen3-Coder-30B-A3B-Instruct: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2025-08-07).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"672e747c77e99032","after":"7e0648e7a298d820"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"703eb35446d63a053064b2812c37c385b41f269f","toSha":"8110d98ad896c995124f3bc44b2f8dd695a4d252","commit":"https://huggingface.co/Qwen/Qwen3-Coder-30B-A3B-Instruct/commit/8110d98ad896c995124f3bc44b2f8dd695a4d252","source":"backfill"},{"id":"Qwen/Qwen3-Coder-480B-A35B-Instruct-FP8","date":"2025-08-07T07:07:09.000Z","classes":["template"],"sentence":"Qwen3-Coder-480B-A35B-Instruct-FP8: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2025-08-07).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"672e747c77e99032","after":"7e0648e7a298d820"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"8fe13515933c3065538cad7de50bbca4b1af3e3b","toSha":"1f6f99b0d0f8e209a7574a946b440b1020a6f31b","commit":"https://huggingface.co/Qwen/Qwen3-Coder-480B-A35B-Instruct-FP8/commit/1f6f99b0d0f8e209a7574a946b440b1020a6f31b","source":"backfill"},{"id":"Qwen/Qwen3-Coder-30B-A3B-Instruct-FP8","date":"2025-08-07T07:04:49.000Z","classes":["template"],"sentence":"Qwen3-Coder-30B-A3B-Instruct-FP8: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2025-08-07).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"672e747c77e99032","after":"7e0648e7a298d820"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"d8df4ca2367d9fa65e4cefc0ef44fcc6ed2f6565","toSha":"acc0a21828d8d3483ddd1ce54d2e98a89ca80309","commit":"https://huggingface.co/Qwen/Qwen3-Coder-30B-A3B-Instruct-FP8/commit/acc0a21828d8d3483ddd1ce54d2e98a89ca80309","source":"backfill"},{"id":"unsloth/gpt-oss-20b-GGUF","date":"2025-08-06T03:50:16.000Z","classes":["structural","metadata"],"sentence":"gpt-oss-20b-GGUF: declared dtype unset to \"bfloat16\"; quantization_config changed; the stored format of the weights is different (2025-08-06).","keys":[{"file":"config","key":"dtype","before":null,"after":"bfloat16"},{"file":"config","key":"quantization_config","before":{"quant_method":"mxfp4","_rest":"ca7acd1fa0c4a5b3"},"after":null}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"4782614e97a6ca8b27ebf9c8916ac15f0a850887","toSha":"5e4ccdfff3117bf112e08ff7e024ed55447e8e67","commit":"https://huggingface.co/unsloth/gpt-oss-20b-GGUF/commit/5e4ccdfff3117bf112e08ff7e024ed55447e8e67","source":"backfill"},{"id":"unsloth/gpt-oss-20b-BF16","date":"2025-08-05T22:00:47.000Z","classes":["template"],"sentence":"gpt-oss-20b-BF16: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2025-08-05).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"f8d9255777615591","after":"445c3a7c29d9cf61"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"570c85e14e8b20677532d61251bc3ba97e71c09a","toSha":"cc89b3e7fd423253264883a80a4fa5abc619649f","commit":"https://huggingface.co/unsloth/gpt-oss-20b-BF16/commit/cc89b3e7fd423253264883a80a4fa5abc619649f","source":"backfill"},{"id":"openai/gpt-oss-120b","date":"2025-08-05T15:45:04.000Z","classes":["template"],"sentence":"gpt-oss-120b: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2025-08-05).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"3aff18eca7ee88d4","after":"f8d9255777615591"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"8c0580383cb1e6a9157669336ade6797a024cd9a","toSha":"f39accdb091b6c5710dbbf02674c27064e35019b","commit":"https://huggingface.co/openai/gpt-oss-120b/commit/f39accdb091b6c5710dbbf02674c27064e35019b","source":"backfill"},{"id":"openai/gpt-oss-20b","date":"2025-08-05T15:40:37.000Z","classes":["template"],"sentence":"gpt-oss-20b: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2025-08-05).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"3aff18eca7ee88d4","after":"f8d9255777615591"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"f81fef1ddd90d214968e951a76834f1ded130a18","toSha":"ec854da5735f125fe36f080d8013482590f9ad7d","commit":"https://huggingface.co/openai/gpt-oss-20b/commit/ec854da5735f125fe36f080d8013482590f9ad7d","source":"backfill"},{"id":"Qwen/Qwen3-Coder-30B-A3B-Instruct-FP8","date":"2025-07-31T08:30:02.000Z","classes":["tokenizer","generation-defaults"],"sentence":"Qwen3-Coder-30B-A3B-Instruct-FP8: EOS changed from 151645 to [151645,151643]; anything fine-tuned or served against the old value stops at a different token; pad token unset to 151643; batched training masks a different id; generation defaults changed (do_sample unset to true, repetition_penalty unset to 1.05, temperature unset to 0.7, top_k unset to 20); a call that does not pass these explicitly now generates differently (2025-07-31).","keys":[{"file":"generation","key":"do_sample","before":null,"after":true},{"file":"generation","key":"eos_token_id","before":151645,"after":[151645,151643]},{"file":"generation","key":"pad_token_id","before":null,"after":151643},{"file":"generation","key":"repetition_penalty","before":null,"after":1.05},{"file":"generation","key":"temperature","before":null,"after":0.7},{"file":"generation","key":"top_k","before":null,"after":20},{"file":"generation","key":"top_p","before":null,"after":0.8}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":{"before":151645,"after":[151645,151643]},"bos":null},"fromSha":"3e3d2d849dab9d323d2d0355604bb54e20547be0","toSha":"d8df4ca2367d9fa65e4cefc0ef44fcc6ed2f6565","commit":"https://huggingface.co/Qwen/Qwen3-Coder-30B-A3B-Instruct-FP8/commit/d8df4ca2367d9fa65e4cefc0ef44fcc6ed2f6565","source":"backfill"},{"id":"Qwen/Qwen3-30B-A3B-Instruct-2507","date":"2025-07-29T12:38:27.000Z","classes":["generation-defaults"],"sentence":"Qwen3-30B-A3B-Instruct-2507: generation defaults changed (temperature 0.6 to 0.7, top_p 0.95 to 0.8); a call that does not pass these explicitly now generates differently (2025-07-29).","keys":[{"file":"generation","key":"temperature","before":0.6,"after":0.7},{"file":"generation","key":"top_p","before":0.95,"after":0.8}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"fd5492ea24611e203da0b593be7916913e490362","toSha":"97f3f7d6f068582355b6fb44014a1b60e282636b","commit":"https://huggingface.co/Qwen/Qwen3-30B-A3B-Instruct-2507/commit/97f3f7d6f068582355b6fb44014a1b60e282636b","source":"backfill"},{"id":"Qwen/Qwen3-30B-A3B-Instruct-2507-FP8","date":"2025-07-29T12:37:39.000Z","classes":["tokenizer","template"],"sentence":"Qwen3-30B-A3B-Instruct-2507-FP8: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering; tokenizer model_max_length 131072 to 262144; truncation defaults change (2025-07-29).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"a55ee1b1660128b7","after":"40c21f34cf67d8c7"},{"file":"tokenizer","key":"model_max_length","before":131072,"after":262144}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"2d356c75c5f4a4c194960771ad6b7ffbdf451b22","toSha":"81e34e368e9e39d3e6967d27fef73607a3ed96f0","commit":"https://huggingface.co/Qwen/Qwen3-30B-A3B-Instruct-2507-FP8/commit/81e34e368e9e39d3e6967d27fef73607a3ed96f0","source":"backfill"},{"id":"Qwen/Qwen3-30B-A3B-Instruct-2507","date":"2025-07-29T12:11:03.000Z","classes":["tokenizer","template"],"sentence":"Qwen3-30B-A3B-Instruct-2507: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering; tokenizer model_max_length 131072 to 262144; truncation defaults change (2025-07-29).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"a55ee1b1660128b7","after":"40c21f34cf67d8c7"},{"file":"tokenizer","key":"model_max_length","before":131072,"after":262144}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"023ad300e0cf991b9e6aeacade6e7bc71185fe71","toSha":"fd5492ea24611e203da0b593be7916913e490362","commit":"https://huggingface.co/Qwen/Qwen3-30B-A3B-Instruct-2507/commit/fd5492ea24611e203da0b593be7916913e490362","source":"backfill"},{"id":"Qwen/Qwen3-30B-A3B-Instruct-2507","date":"2025-07-29T10:44:37.000Z","classes":["generation-defaults"],"sentence":"Qwen3-30B-A3B-Instruct-2507: generation defaults changed (temperature 0.7 to 0.6, top_p 0.8 to 0.95); a call that does not pass these explicitly now generates differently (2025-07-29).","keys":[{"file":"generation","key":"temperature","before":0.7,"after":0.6},{"file":"generation","key":"top_p","before":0.8,"after":0.95}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"b92fe559a97985763a05d1965db70c2f7e353b78","toSha":"023ad300e0cf991b9e6aeacade6e7bc71185fe71","commit":"https://huggingface.co/Qwen/Qwen3-30B-A3B-Instruct-2507/commit/023ad300e0cf991b9e6aeacade6e7bc71185fe71","source":"backfill"},{"id":"Qwen/Qwen3-30B-A3B-Instruct-2507-FP8","date":"2025-07-29T01:59:38.000Z","classes":["tokenizer","generation-defaults"],"sentence":"Qwen3-30B-A3B-Instruct-2507-FP8: EOS changed from 151645 to [151645,151643]; anything fine-tuned or served against the old value stops at a different token; pad token unset to 151643; batched training masks a different id; generation defaults changed (do_sample unset to true, temperature unset to 0.7, top_k unset to 20, top_p unset to 0.8); a call that does not pass these explicitly now generates differently (2025-07-29).","keys":[{"file":"generation","key":"do_sample","before":null,"after":true},{"file":"generation","key":"eos_token_id","before":151645,"after":[151645,151643]},{"file":"generation","key":"pad_token_id","before":null,"after":151643},{"file":"generation","key":"temperature","before":null,"after":0.7},{"file":"generation","key":"top_k","before":null,"after":20},{"file":"generation","key":"top_p","before":null,"after":0.8}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":{"before":151645,"after":[151645,151643]},"bos":null},"fromSha":"61c2ee12e2e018cd6cdcaa3713e6ffd19e7b8086","toSha":"2d356c75c5f4a4c194960771ad6b7ffbdf451b22","commit":"https://huggingface.co/Qwen/Qwen3-30B-A3B-Instruct-2507-FP8/commit/2d356c75c5f4a4c194960771ad6b7ffbdf451b22","source":"backfill"},{"id":"HuggingFaceTB/SmolLM3-3B","date":"2025-07-28T14:47:10.000Z","classes":["structural"],"sentence":"SmolLM3-3B: tie_word_embeddings unset to true; the output head is now shared with the embedding (2025-07-28).","keys":[{"file":"config","key":"tie_word_embeddings","before":null,"after":true}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"486a602f60b9b61d727ee978b3747824cb659f4d","toSha":"988f1dde6bd2a36e3e7f45777a36507ea004fe22","commit":"https://huggingface.co/HuggingFaceTB/SmolLM3-3B/commit/988f1dde6bd2a36e3e7f45777a36507ea004fe22","source":"backfill"},{"id":"Qwen/Qwen3-30B-A3B-Instruct-2507-FP8","date":"2025-07-28T13:34:23.000Z","classes":["structural"],"sentence":"Qwen3-30B-A3B-Instruct-2507-FP8: MLP width 5504 to 6144; context window 1048576 to 262144 tokens; quantization_config changed; the stored format of the weights is different (2025-07-28).","keys":[{"file":"config","key":"intermediate_size","before":5504,"after":6144},{"file":"config","key":"max_position_embeddings","before":1048576,"after":262144},{"file":"config","key":"quantization_config","before":{"activation_scheme":"dynamic","quant_method":"fp8","_rest":"81cc3297a7cc5d80"},"after":{"activation_scheme":"dynamic","fmt":"e4m3","quant_method":"fp8","_rest":"81cc3297a7cc5d80"}}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":{"before":1048576,"after":262144},"eos":null,"bos":null},"fromSha":"b1d5266ce8b61c86602bd1f2f61520a37288cda7","toSha":"61c2ee12e2e018cd6cdcaa3713e6ffd19e7b8086","commit":"https://huggingface.co/Qwen/Qwen3-30B-A3B-Instruct-2507-FP8/commit/61c2ee12e2e018cd6cdcaa3713e6ffd19e7b8086","source":"backfill"},{"id":"Qwen/Qwen3-30B-A3B-Instruct-2507","date":"2025-07-28T12:57:43.000Z","classes":["structural"],"sentence":"Qwen3-30B-A3B-Instruct-2507: MLP width 5504 to 6144; context window 1048576 to 262144 tokens; architectures [\"Qwen2MoeForCausalLM\"] to [\"Qwen3MoeForCausalLM\"]; a different class is instantiated from the same id (2025-07-28).","keys":[{"file":"config","key":"architectures","before":["Qwen2MoeForCausalLM"],"after":["Qwen3MoeForCausalLM"]},{"file":"config","key":"intermediate_size","before":5504,"after":6144},{"file":"config","key":"max_position_embeddings","before":1048576,"after":262144},{"file":"config","key":"model_type","before":"qwen2_moe","after":"qwen3_moe"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":{"before":1048576,"after":262144},"eos":null,"bos":null},"fromSha":"4457916f69bf6a9a24dfebd20608f00aeef2f349","toSha":"e67ac5d4218aa2fe8b3f095a2128e61f509f2511","commit":"https://huggingface.co/Qwen/Qwen3-30B-A3B-Instruct-2507/commit/e67ac5d4218aa2fe8b3f095a2128e61f509f2511","source":"backfill"},{"id":"Qwen/Qwen3-Coder-480B-A35B-Instruct-FP8","date":"2025-07-22T17:26:20.000Z","classes":["tokenizer","template","generation-defaults"],"sentence":"Qwen3-Coder-480B-A35B-Instruct-FP8: EOS changed from 151645 to [151645,151643]; anything fine-tuned or served against the old value stops at a different token; BOS 151643 removed; pad token unset to 151643; batched training masks a different id (2025-07-22).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":null,"after":"672e747c77e99032"},{"file":"generation","key":"bos_token_id","before":151643,"after":null},{"file":"generation","key":"do_sample","before":null,"after":true},{"file":"generation","key":"eos_token_id","before":151645,"after":[151645,151643]},{"file":"generation","key":"pad_token_id","before":null,"after":151643},{"file":"generation","key":"repetition_penalty","before":null,"after":1.05},{"file":"generation","key":"temperature","before":null,"after":0.7},{"file":"generation","key":"top_k","before":null,"after":20},{"file":"generation","key":"top_p","before":null,"after":0.8}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":{"before":151645,"after":[151645,151643]},"bos":{"before":151643,"after":null}},"fromSha":"64ac916bdfc2707436155454932145f9dcdafdd5","toSha":"8fe13515933c3065538cad7de50bbca4b1af3e3b","commit":"https://huggingface.co/Qwen/Qwen3-Coder-480B-A35B-Instruct-FP8/commit/8fe13515933c3065538cad7de50bbca4b1af3e3b","source":"backfill"},{"id":"hmellor/Ilama-3.2-1B","date":"2025-07-22T15:02:56.000Z","classes":["structural"],"sentence":"Ilama-3.2-1B: architectures [\"LlamaForCausalLM\"] to [\"IlamaForCausalLM\"]; a different class is instantiated from the same id (2025-07-22).","keys":[{"file":"config","key":"architectures","before":["LlamaForCausalLM"],"after":["IlamaForCausalLM"]},{"file":"config","key":"model_type","before":"llama","after":"ilama"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"b04059ced924fcd2f635cc23867e740753724b4f","toSha":"923c85f0b1c189d68905dd2f6949590b2a5d2016","commit":"https://huggingface.co/hmellor/Ilama-3.2-1B/commit/923c85f0b1c189d68905dd2f6949590b2a5d2016","source":"backfill"},{"id":"Qwen/Qwen3-235B-A22B-Instruct-2507-FP8","date":"2025-07-21T11:58:32.000Z","classes":["tokenizer","generation-defaults"],"sentence":"Qwen3-235B-A22B-Instruct-2507-FP8: EOS changed from 151645 to [151645,151643]; anything fine-tuned or served against the old value stops at a different token; pad token unset to 151643; batched training masks a different id; generation defaults changed (do_sample unset to true, temperature unset to 0.7, top_k unset to 20, top_p unset to 0.8); a call that does not pass these explicitly now generates differently (2025-07-21).","keys":[{"file":"generation","key":"do_sample","before":null,"after":true},{"file":"generation","key":"eos_token_id","before":151645,"after":[151645,151643]},{"file":"generation","key":"pad_token_id","before":null,"after":151643},{"file":"generation","key":"temperature","before":null,"after":0.7},{"file":"generation","key":"top_k","before":null,"after":20},{"file":"generation","key":"top_p","before":null,"after":0.8}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":{"before":151645,"after":[151645,151643]},"bos":null},"fromSha":"c581b8ffb8fd4b5d527fbb87508b36d2a70fdcb1","toSha":"652b9036a3d93443261dfe5ed47192ba859f14d3","commit":"https://huggingface.co/Qwen/Qwen3-235B-A22B-Instruct-2507-FP8/commit/652b9036a3d93443261dfe5ed47192ba859f14d3","source":"backfill"},{"id":"Qwen/Qwen3-235B-A22B-Instruct-2507-FP8","date":"2025-07-21T11:57:34.000Z","classes":["tokenizer","template"],"sentence":"Qwen3-235B-A22B-Instruct-2507-FP8: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering; tokenizer model_max_length 131072 to 262144; truncation defaults change (2025-07-21).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"a55ee1b1660128b7","after":"40c21f34cf67d8c7"},{"file":"tokenizer","key":"model_max_length","before":131072,"after":262144}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"b939f17ff56df5c734de72b3d06cb51164ac0489","toSha":"c581b8ffb8fd4b5d527fbb87508b36d2a70fdcb1","commit":"https://huggingface.co/Qwen/Qwen3-235B-A22B-Instruct-2507-FP8/commit/c581b8ffb8fd4b5d527fbb87508b36d2a70fdcb1","source":"backfill"},{"id":"Qwen/Qwen3-235B-A22B-Instruct-2507-FP8","date":"2025-07-21T11:54:29.000Z","classes":["structural"],"sentence":"Qwen3-235B-A22B-Instruct-2507-FP8: MLP width 10944 to 12288; context window 1048576 to 262144 tokens; quantization_config changed; the stored format of the weights is different (2025-07-21).","keys":[{"file":"config","key":"intermediate_size","before":10944,"after":12288},{"file":"config","key":"max_position_embeddings","before":1048576,"after":262144},{"file":"config","key":"quantization_config","before":{"activation_scheme":"dynamic","quant_method":"fp8","_rest":"fc843cfb0186718b"},"after":{"activation_scheme":"dynamic","fmt":"e4m3","quant_method":"fp8","_rest":"fc843cfb0186718b"}}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":{"before":1048576,"after":262144},"eos":null,"bos":null},"fromSha":"d29dc93a6c1520d3bd98304d16ae7834a1e1ef69","toSha":"b939f17ff56df5c734de72b3d06cb51164ac0489","commit":"https://huggingface.co/Qwen/Qwen3-235B-A22B-Instruct-2507-FP8/commit/b939f17ff56df5c734de72b3d06cb51164ac0489","source":"backfill"},{"id":"moonshotai/Kimi-K2-Instruct","date":"2025-07-18T06:28:42.000Z","classes":["template"],"sentence":"Kimi-K2-Instruct: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2025-07-18).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"aab20feb9bc6881f","after":"454036503ee2fc14"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"37a95d86572ddbd940d06059c00b138b2098a5dc","toSha":"6a1f5e6be5b1ddfd220836faed223dd70fbc091e","commit":"https://huggingface.co/moonshotai/Kimi-K2-Instruct/commit/6a1f5e6be5b1ddfd220836faed223dd70fbc091e","source":"backfill"},{"id":"moonshotai/Kimi-K2-Instruct","date":"2025-07-15T07:09:19.000Z","classes":["template"],"sentence":"Kimi-K2-Instruct: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2025-07-15).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"d33ebe8701707e7b","after":"aab20feb9bc6881f"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"1221eb6b4da3ef69743a391ae7c314ea10918a0d","toSha":"37a95d86572ddbd940d06059c00b138b2098a5dc","commit":"https://huggingface.co/moonshotai/Kimi-K2-Instruct/commit/37a95d86572ddbd940d06059c00b138b2098a5dc","source":"backfill"},{"id":"moonshotai/Kimi-K2-Instruct","date":"2025-07-15T07:07:05.000Z","classes":["template"],"sentence":"Kimi-K2-Instruct: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2025-07-15).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"0d19c3a601c02d8f","after":"d33ebe8701707e7b"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"22bf139598edbbbcd9226c6ea392200d4b53f4b6","toSha":"1221eb6b4da3ef69743a391ae7c314ea10918a0d","commit":"https://huggingface.co/moonshotai/Kimi-K2-Instruct/commit/1221eb6b4da3ef69743a391ae7c314ea10918a0d","source":"backfill"},{"id":"HuggingFaceTB/SmolLM3-3B","date":"2025-07-09T10:05:39.000Z","classes":["template"],"sentence":"SmolLM3-3B: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2025-07-09).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"0dab472abbdc499c","after":"3e0e275a527f2496"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"a169fa445c7eb1c4df03da97cd914dbdb9a94421","toSha":"486a602f60b9b61d727ee978b3747824cb659f4d","commit":"https://huggingface.co/HuggingFaceTB/SmolLM3-3B/commit/486a602f60b9b61d727ee978b3747824cb659f4d","source":"backfill"},{"id":"HuggingFaceTB/SmolLM3-3B","date":"2025-07-08T14:28:27.000Z","classes":["template"],"sentence":"SmolLM3-3B: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2025-07-08).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"af3a186af912d9e1","after":"0dab472abbdc499c"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"2ea4d77711467794bfae4d525e1995b32191ce4d","toSha":"a169fa445c7eb1c4df03da97cd914dbdb9a94421","commit":"https://huggingface.co/HuggingFaceTB/SmolLM3-3B/commit/a169fa445c7eb1c4df03da97cd914dbdb9a94421","source":"backfill"},{"id":"HuggingFaceTB/SmolLM3-3B","date":"2025-07-08T14:23:54.000Z","classes":["generation-defaults"],"sentence":"SmolLM3-3B: generation defaults changed (do_sample unset to true, temperature unset to 0.6, top_p unset to 0.95); a call that does not pass these explicitly now generates differently (2025-07-08).","keys":[{"file":"generation","key":"do_sample","before":null,"after":true},{"file":"generation","key":"temperature","before":null,"after":0.6},{"file":"generation","key":"top_p","before":null,"after":0.95}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"82547f52c3a407c66c899145a29e4281e4c11e26","toSha":"2ea4d77711467794bfae4d525e1995b32191ce4d","commit":"https://huggingface.co/HuggingFaceTB/SmolLM3-3B/commit/2ea4d77711467794bfae4d525e1995b32191ce4d","source":"backfill"},{"id":"HuggingFaceTB/SmolLM3-3B","date":"2025-07-08T10:15:06.000Z","classes":["tokenizer"],"sentence":"SmolLM3-3B: EOS changed from 128001 to 128012; anything fine-tuned or served against the old value stops at a different token (2025-07-08).","keys":[{"file":"config","key":"eos_token_id","before":128001,"after":128012}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":{"before":128001,"after":128012},"bos":null},"fromSha":"451a3a6152b50e3c22673dfd6bc512aa7094054f","toSha":"82547f52c3a407c66c899145a29e4281e4c11e26","commit":"https://huggingface.co/HuggingFaceTB/SmolLM3-3B/commit/82547f52c3a407c66c899145a29e4281e4c11e26","source":"backfill"},{"id":"jinaai/jina-embeddings-v4","date":"2025-07-03T14:30:30.000Z","classes":["metadata"],"sentence":"jina-embeddings-v4: derived parameters 2.80B to 3.09B (+288.2M, +0.58 GB of bf16 weights); model_type declared as \"qwen2_5_vl_text\" (was unset) (2025-07-03).","keys":[{"file":"config","key":"model_type","before":null,"after":"qwen2_5_vl_text"}],"consequence":{"params":{"before":2801396736,"after":3089550848},"weightsGbBf16":{"before":5.603,"after":6.179},"trainPeakGb":{"before":253.987,"after":356.717},"fingerprint":{"before":"be837245","after":"2db78954"},"context":null,"eos":null,"bos":null},"fromSha":"d406962fb3674781aa5afb509e24a67d20a9e33a","toSha":"e8ce774a3630bebaa0c594c6833b720fbb3af8c2","commit":"https://huggingface.co/jinaai/jina-embeddings-v4/commit/e8ce774a3630bebaa0c594c6833b720fbb3af8c2","source":"backfill"},{"id":"amd/DeepSeek-R1-MXFP4","date":"2025-06-12T17:13:38.000Z","classes":["tokenizer"],"sentence":"DeepSeek-R1-MXFP4: tokenizer model_max_length 512 to 16384; truncation defaults change (2025-06-12).","keys":[{"file":"tokenizer","key":"model_max_length","before":512,"after":16384}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"5e05bd6b1c5404480f9743f6606cce0b08bee923","toSha":"a8b84c8ab78875ec0b8a5c15c6bf683b3cc50a53","commit":"https://huggingface.co/amd/DeepSeek-R1-MXFP4/commit/a8b84c8ab78875ec0b8a5c15c6bf683b3cc50a53","source":"backfill"},{"id":"ai-sage/Giga-Embeddings-instruct","date":"2025-06-04T09:25:41.000Z","classes":["structural"],"sentence":"Giga-Embeddings-instruct: context window 32768 to 4096 tokens (2025-06-04).","keys":[{"file":"config","key":"max_position_embeddings","before":32768,"after":4096}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":{"before":32768,"after":4096},"eos":null,"bos":null},"fromSha":"0cd92734c456a6571957a2ade0573c57faf996f9","toSha":"f4891741cf49fc2e724a14188470cd57c39fec93","commit":"https://huggingface.co/ai-sage/Giga-Embeddings-instruct/commit/f4891741cf49fc2e724a14188470cd57c39fec93","source":"backfill"},{"id":"Qwen/Qwen3-Embedding-0.6B","date":"2025-06-03T14:30:44.000Z","classes":["structural","tokenizer","generation-defaults"],"sentence":"Qwen3-Embedding-0.6B: EOS changed from [151645,151643] to 151643; anything fine-tuned or served against the old value stops at a different token; context window 40960 to 32768 tokens; pad token 151643 to unset; batched training masks a different id (2025-06-03).","keys":[{"file":"config","key":"eos_token_id","before":151645,"after":151643},{"file":"config","key":"max_position_embeddings","before":40960,"after":32768},{"file":"generation","key":"do_sample","before":true,"after":null},{"file":"generation","key":"eos_token_id","before":[151645,151643],"after":151643},{"file":"generation","key":"max_new_tokens","before":null,"after":2048},{"file":"generation","key":"pad_token_id","before":151643,"after":null},{"file":"generation","key":"temperature","before":0.6,"after":null},{"file":"generation","key":"top_k","before":20,"after":null},{"file":"generation","key":"top_p","before":0.95,"after":null}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":{"before":40960,"after":32768},"eos":{"before":[151645,151643],"after":151643},"bos":null},"fromSha":"1c68a8439911c7a3d16e68f03bec384ee4bf3b81","toSha":"99cabfa1346cbf4ac8b0e73079bb2e286cff3a1f","commit":"https://huggingface.co/Qwen/Qwen3-Embedding-0.6B/commit/99cabfa1346cbf4ac8b0e73079bb2e286cff3a1f","source":"backfill"},{"id":"deepseek-ai/DeepSeek-R1-0528-Qwen3-8B","date":"2025-05-29T13:13:34.000Z","classes":["structural"],"sentence":"DeepSeek-R1-0528-Qwen3-8B: rope_scaling rewritten in a different format ({\"factor\":4,\"original_max_position_embeddings\":32768,\"rop... to {\"attn_factor\":0.8782488562869419,\"factor\":4,\"original_ma...); a loader that only reads the old format rejects or ignores the new one (2025-05-29).","keys":[{"file":"config","key":"rope_scaling","before":{"rope_type":"yarn","factor":4,"original_max_position_embeddings":32768},"after":{"rope_type":"yarn","factor":4,"original_max_position_embeddings":32768,"attn_factor":0.8782488562869419}}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"767459076260d2a2f07718f494cbc63034913e80","toSha":"6e8885a6ff5c1dc5201574c8fd700323f23c25fa","commit":"https://huggingface.co/deepseek-ai/DeepSeek-R1-0528-Qwen3-8B/commit/6e8885a6ff5c1dc5201574c8fd700323f23c25fa","source":"backfill"},{"id":"ai-sage/Giga-Embeddings-instruct","date":"2025-05-29T12:24:58.000Z","classes":["structural"],"sentence":"Giga-Embeddings-instruct: layer count 27 to 36; derived parameters 2.61B to 3.30B (+693.7M, +1.39 GB of bf16 weights) (2025-05-29).","keys":[{"file":"config","key":"num_hidden_layers","before":27,"after":36}],"consequence":{"params":{"before":2606476544,"after":3300146432},"weightsGbBf16":{"before":5.213,"after":6.6},"trainPeakGb":{"before":296.817,"after":375.903},"fingerprint":{"before":"42de84f1","after":"87540c74"},"context":null,"eos":null,"bos":null},"fromSha":"201baa9d6f8f22c095d09b61fb50228d9382306f","toSha":"e5a07187e0d1e63d4faca11a658bec40e1ccee5c","commit":"https://huggingface.co/ai-sage/Giga-Embeddings-instruct/commit/e5a07187e0d1e63d4faca11a658bec40e1ccee5c","source":"backfill"},{"id":"deepseek-ai/DeepSeek-R1-0528","date":"2025-05-29T11:37:44.000Z","classes":["template"],"sentence":"DeepSeek-R1-0528: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2025-05-29).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"3a3036eeb96ca0e4","after":"53671bac29bcb3e1"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"11628360bdbb84a195bb216d98bc724f6af08d57","toSha":"4236a6af538feda4548eca9ab308586007567f52","commit":"https://huggingface.co/deepseek-ai/DeepSeek-R1-0528/commit/4236a6af538feda4548eca9ab308586007567f52","source":"backfill"},{"id":"Qwen/Qwen3-8B-Base","date":"2025-05-20T12:05:22.000Z","classes":["tokenizer"],"sentence":"Qwen3-8B-Base: EOS changed from \"<|im_end|>\" to \"<|endoftext|>\"; anything fine-tuned or served against the old value stops at a different token (2025-05-20).","keys":[{"file":"tokenizer","key":"eos_token","before":"<|im_end|>","after":"<|endoftext|>"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":{"before":"<|im_end|>","after":"<|endoftext|>"},"bos":null},"fromSha":"a64df03b50f8fdeb56e303ed2dd5228dc149f926","toSha":"b86f722ec0768e8849ddeb2d8ce3a518bf4fd504","commit":"https://huggingface.co/Qwen/Qwen3-8B-Base/commit/b86f722ec0768e8849ddeb2d8ce3a518bf4fd504","source":"backfill"},{"id":"Qwen/Qwen3-4B-Base","date":"2025-05-20T12:04:42.000Z","classes":["tokenizer"],"sentence":"Qwen3-4B-Base: EOS changed from \"<|im_end|>\" to \"<|endoftext|>\"; anything fine-tuned or served against the old value stops at a different token (2025-05-20).","keys":[{"file":"tokenizer","key":"eos_token","before":"<|im_end|>","after":"<|endoftext|>"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":{"before":"<|im_end|>","after":"<|endoftext|>"},"bos":null},"fromSha":"08a151eb7f61cd599444a3fb2fe2008762111300","toSha":"08238297d7a5db40bf8d7513ce96da39f41d3b17","commit":"https://huggingface.co/Qwen/Qwen3-4B-Base/commit/08238297d7a5db40bf8d7513ce96da39f41d3b17","source":"backfill"},{"id":"Qwen/Qwen3-1.7B-Base","date":"2025-05-20T12:03:42.000Z","classes":["tokenizer"],"sentence":"Qwen3-1.7B-Base: EOS changed from \"<|im_end|>\" to \"<|endoftext|>\"; anything fine-tuned or served against the old value stops at a different token (2025-05-20).","keys":[{"file":"tokenizer","key":"eos_token","before":"<|im_end|>","after":"<|endoftext|>"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":{"before":"<|im_end|>","after":"<|endoftext|>"},"bos":null},"fromSha":"36be17a0ee54955c2d50eb4af5a126b429874a6e","toSha":"912d2727784ca0a6f718845aa14d4d9e5f48fe26","commit":"https://huggingface.co/Qwen/Qwen3-1.7B-Base/commit/912d2727784ca0a6f718845aa14d4d9e5f48fe26","source":"backfill"},{"id":"Qwen/Qwen3-0.6B-Base","date":"2025-05-20T12:03:21.000Z","classes":["tokenizer"],"sentence":"Qwen3-0.6B-Base: EOS changed from \"<|im_end|>\" to \"<|endoftext|>\"; anything fine-tuned or served against the old value stops at a different token (2025-05-20).","keys":[{"file":"tokenizer","key":"eos_token","before":"<|im_end|>","after":"<|endoftext|>"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":{"before":"<|im_end|>","after":"<|endoftext|>"},"bos":null},"fromSha":"527f918d05b5c4bcfc25c13b054f06ff8e906e82","toSha":"311c62e88814bff7206909ccd330bab0a784743b","commit":"https://huggingface.co/Qwen/Qwen3-0.6B-Base/commit/311c62e88814bff7206909ccd330bab0a784743b","source":"backfill"},{"id":"Qwen/Qwen3-8B-FP8","date":"2025-05-19T07:46:45.000Z","classes":["template"],"sentence":"Qwen3-8B-FP8: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2025-05-19).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"87a2728cb8dc9fe4","after":"a55ee1b1660128b7"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"add3e8dee80bcfc07aca2fd9edb299ae5c1cb8f7","toSha":"a8f74dfface525a7143ff3a196248434898fdf04","commit":"https://huggingface.co/Qwen/Qwen3-8B-FP8/commit/a8f74dfface525a7143ff3a196248434898fdf04","source":"backfill"},{"id":"Qwen/Qwen3-8B-AWQ","date":"2025-05-19T07:46:28.000Z","classes":["template"],"sentence":"Qwen3-8B-AWQ: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2025-05-19).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"87a2728cb8dc9fe4","after":"a55ee1b1660128b7"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"9e07647a76edf7424dfe621814b6a1fe8fc54d4a","toSha":"4a2d74b63371c8cd724aa15dc5efbd329b16f573","commit":"https://huggingface.co/Qwen/Qwen3-8B-AWQ/commit/4a2d74b63371c8cd724aa15dc5efbd329b16f573","source":"backfill"},{"id":"Qwen/Qwen3-8B","date":"2025-05-19T07:46:12.000Z","classes":["template"],"sentence":"Qwen3-8B: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2025-05-19).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"87a2728cb8dc9fe4","after":"a55ee1b1660128b7"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"47719a242beab8f9aecc40ce3928b034dd5dd559","toSha":"895c8d171bc03c30e113cd7a28c02494b5e068b7","commit":"https://huggingface.co/Qwen/Qwen3-8B/commit/895c8d171bc03c30e113cd7a28c02494b5e068b7","source":"backfill"},{"id":"Qwen/Qwen3-4B","date":"2025-05-19T07:44:39.000Z","classes":["template"],"sentence":"Qwen3-4B: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2025-05-19).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"87a2728cb8dc9fe4","after":"a55ee1b1660128b7"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"3101254bbe4169895668a0e7653c3fd1f313576e","toSha":"8136a03248eac0530d0f630f9a4e21810f9137de","commit":"https://huggingface.co/Qwen/Qwen3-4B/commit/8136a03248eac0530d0f630f9a4e21810f9137de","source":"backfill"},{"id":"Qwen/Qwen3-32B-AWQ","date":"2025-05-19T07:44:10.000Z","classes":["template"],"sentence":"Qwen3-32B-AWQ: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2025-05-19).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"87a2728cb8dc9fe4","after":"a55ee1b1660128b7"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"d5f8dfea94a4df504de08bd3670cd5fdfb6a92ce","toSha":"1e09580a8dacf7685baf956d86ca7a13c7506910","commit":"https://huggingface.co/Qwen/Qwen3-32B-AWQ/commit/1e09580a8dacf7685baf956d86ca7a13c7506910","source":"backfill"},{"id":"Qwen/Qwen3-32B","date":"2025-05-19T07:43:51.000Z","classes":["template"],"sentence":"Qwen3-32B: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2025-05-19).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"87a2728cb8dc9fe4","after":"a55ee1b1660128b7"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"1aecd22f7ae87d957e541aaf933345c33d8a3a26","toSha":"a09d5ccb54a84a572fcde9e54da24ec6e5a4b4a2","commit":"https://huggingface.co/Qwen/Qwen3-32B/commit/a09d5ccb54a84a572fcde9e54da24ec6e5a4b4a2","source":"backfill"},{"id":"Qwen/Qwen3-30B-A3B","date":"2025-05-19T07:42:56.000Z","classes":["template"],"sentence":"Qwen3-30B-A3B: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2025-05-19).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"87a2728cb8dc9fe4","after":"a55ee1b1660128b7"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"d47d535f78ec44bd57128f8e8aeba17eeb0285ea","toSha":"f0abbbee7f4db139f64a2c3be5dc133fd7328acb","commit":"https://huggingface.co/Qwen/Qwen3-30B-A3B/commit/f0abbbee7f4db139f64a2c3be5dc133fd7328acb","source":"backfill"},{"id":"Qwen/Qwen3-235B-A22B","date":"2025-05-19T07:42:02.000Z","classes":["template"],"sentence":"Qwen3-235B-A22B: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2025-05-19).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"87a2728cb8dc9fe4","after":"a55ee1b1660128b7"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"d0aa57eb283a2d2489020470fde653a310b4ec6a","toSha":"d5280577feb14062e43e5d6ad97fcdf9d2f581cc","commit":"https://huggingface.co/Qwen/Qwen3-235B-A22B/commit/d5280577feb14062e43e5d6ad97fcdf9d2f581cc","source":"backfill"},{"id":"Qwen/Qwen3-14B-AWQ","date":"2025-05-19T07:41:24.000Z","classes":["template"],"sentence":"Qwen3-14B-AWQ: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2025-05-19).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"87a2728cb8dc9fe4","after":"a55ee1b1660128b7"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"bac7b87017e34938f468b4ba40b49247c83fa44a","toSha":"1a6fe1ecf891437a270cce11ad54d796c4f56ce0","commit":"https://huggingface.co/Qwen/Qwen3-14B-AWQ/commit/1a6fe1ecf891437a270cce11ad54d796c4f56ce0","source":"backfill"},{"id":"Qwen/Qwen3-14B","date":"2025-05-19T07:41:02.000Z","classes":["template"],"sentence":"Qwen3-14B: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2025-05-19).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"87a2728cb8dc9fe4","after":"a55ee1b1660128b7"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"f8c293defed9c6a8026291e8e7d780697e88d7af","toSha":"ec3b5103878832857b9e0f6ab32eb37ee5e28e9a","commit":"https://huggingface.co/Qwen/Qwen3-14B/commit/ec3b5103878832857b9e0f6ab32eb37ee5e28e9a","source":"backfill"},{"id":"Qwen/Qwen3-1.7B","date":"2025-05-19T07:40:08.000Z","classes":["template"],"sentence":"Qwen3-1.7B: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2025-05-19).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"87a2728cb8dc9fe4","after":"a55ee1b1660128b7"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"8d4744f9e13072f4920c326350fa81eedb74eae9","toSha":"0ecc654158dd093daab47fd37c5f122925d93983","commit":"https://huggingface.co/Qwen/Qwen3-1.7B/commit/0ecc654158dd093daab47fd37c5f122925d93983","source":"backfill"},{"id":"Qwen/Qwen3-0.6B","date":"2025-05-19T07:39:20.000Z","classes":["template"],"sentence":"Qwen3-0.6B: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2025-05-19).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"87a2728cb8dc9fe4","after":"a55ee1b1660128b7"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"16706fc57485378d4ffaf54139b29ccc66ae08fa","toSha":"7e4ae267688d671ddfca3122e4528ee980cf3234","commit":"https://huggingface.co/Qwen/Qwen3-0.6B/commit/7e4ae267688d671ddfca3122e4528ee980cf3234","source":"backfill"},{"id":"jinaai/jina-embeddings-v4","date":"2025-05-16T11:58:33.000Z","classes":["structural"],"sentence":"jina-embeddings-v4: derived parameters 3.09B to 2.80B (-288.3M, -0.58 GB of bf16 weights); architectures [\"ColQwen25Duo\"] to [\"JinaEmbeddingsV4Model\"]; a different class is instantiated from the same id (2025-05-16).","keys":[{"file":"config","key":"architectures","before":["ColQwen25Duo"],"after":["JinaEmbeddingsV4Model"]}],"consequence":{"params":{"before":3089704448,"after":2801396736},"weightsGbBf16":{"before":6.179,"after":5.603},"trainPeakGb":{"before":348.967,"after":253.987},"fingerprint":{"before":"2db78954","after":"be837245"},"context":null,"eos":null,"bos":null},"fromSha":"e3cd201ac75f0648df0b65b89ee97ae5faafc1d0","toSha":"998398d6554c4832d09be1d217794261ac0d5638","commit":"https://huggingface.co/jinaai/jina-embeddings-v4/commit/998398d6554c4832d09be1d217794261ac0d5638","source":"backfill"},{"id":"Qwen/Qwen3-8B-AWQ","date":"2025-05-08T05:06:14.000Z","classes":["structural"],"sentence":"Qwen3-8B-AWQ: declared dtype \"bfloat16\" to \"float16\" (2025-05-08).","keys":[{"file":"config","key":"dtype","before":"bfloat16","after":"float16"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"c664f5e955c33186a8b7a5cf0cbc3b220e4afdd4","toSha":"9e07647a76edf7424dfe621814b6a1fe8fc54d4a","commit":"https://huggingface.co/Qwen/Qwen3-8B-AWQ/commit/9e07647a76edf7424dfe621814b6a1fe8fc54d4a","source":"backfill"},{"id":"fixie-ai/ultravox-v0_5-llama-3_2-1b","date":"2025-05-06T20:31:24.000Z","classes":["tokenizer"],"sentence":"ultravox-v0_5-llama-3_2-1b: tokenizer_class \"PreTrainedTokenizerFast\" to \"PreTrainedTokenizer\" (2025-05-06).","keys":[{"file":"tokenizer","key":"tokenizer_class","before":"PreTrainedTokenizerFast","after":"PreTrainedTokenizer"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"de683ba4428d3795e7a26d4c0277afbb7507c31c","toSha":"69c6e87e9653b9b396b501fcf20abfd37da42995","commit":"https://huggingface.co/fixie-ai/ultravox-v0_5-llama-3_2-1b/commit/69c6e87e9653b9b396b501fcf20abfd37da42995","source":"backfill"},{"id":"DavidAU/Qwen3-8B-64k-Context-2X-Josiefied-Uncensored","date":"2025-05-06T06:08:23.000Z","classes":["structural"],"sentence":"Qwen3-8B-64k-Context-2X-Josiefied-Uncensored: context window 40960 to 65536 tokens (2025-05-06).","keys":[{"file":"config","key":"max_position_embeddings","before":40960,"after":65536}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":{"before":40960,"after":65536},"eos":null,"bos":null},"fromSha":"5324f18076c6f884e3d435dd45c408deddbb9b26","toSha":"e12b295987a665749f03dee1dd553bdb9cc23fe5","commit":"https://huggingface.co/DavidAU/Qwen3-8B-64k-Context-2X-Josiefied-Uncensored/commit/e12b295987a665749f03dee1dd553bdb9cc23fe5","source":"backfill"},{"id":"DavidAU/Qwen3-8B-64k-Context-2X-Josiefied-Uncensored","date":"2025-05-06T05:43:21.000Z","classes":["structural"],"sentence":"Qwen3-8B-64k-Context-2X-Josiefied-Uncensored: rope_scaling added ({\"factor\":2,\"original_max_position_embeddings\":32768,\"typ...); positions are now scaled (2025-05-06).","keys":[{"file":"config","key":"rope_scaling","before":null,"after":{"type":"yarn","factor":2,"original_max_position_embeddings":32768}}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"980260fabec65c7e64fbe16c8fe98a0b7da5744c","toSha":"5324f18076c6f884e3d435dd45c408deddbb9b26","commit":"https://huggingface.co/DavidAU/Qwen3-8B-64k-Context-2X-Josiefied-Uncensored/commit/5324f18076c6f884e3d435dd45c408deddbb9b26","source":"backfill"},{"id":"Qwen/Qwen3-8B-AWQ","date":"2025-05-03T12:31:05.000Z","classes":["structural","tokenizer","template","generation-defaults"],"sentence":"Qwen3-8B-AWQ: EOS changed from 151645 to [151645,151643]; anything fine-tuned or served against the old value stops at a different token; BOS declared as 151643 (was unset); context window 65536 to 40960 tokens (2025-05-03).","keys":[{"file":"config","key":"bos_token_id","before":null,"after":151643},{"file":"config","key":"max_position_embeddings","before":65536,"after":40960},{"file":"config","key":"pad_token_id","before":151643,"after":null},{"file":"tokenizer","key":"chat_template_sha256","before":"be3e45a0ad8b7079","after":"87a2728cb8dc9fe4"},{"file":"generation","key":"bos_token_id","before":null,"after":151643},{"file":"generation","key":"eos_token_id","before":151645,"after":[151645,151643]},{"file":"generation","key":"temperature","before":null,"after":0.6},{"file":"generation","key":"top_k","before":null,"after":20},{"file":"generation","key":"top_p","before":null,"after":0.95}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":{"before":65536,"after":40960},"eos":{"before":151645,"after":[151645,151643]},"bos":{"before":null,"after":151643}},"fromSha":"cb7d6a337aadb4d2082ed0dcef1032e4f8645194","toSha":"c664f5e955c33186a8b7a5cf0cbc3b220e4afdd4","commit":"https://huggingface.co/Qwen/Qwen3-8B-AWQ/commit/c664f5e955c33186a8b7a5cf0cbc3b220e4afdd4","source":"backfill"},{"id":"Qwen/Qwen3-32B-AWQ","date":"2025-05-01T14:28:51.000Z","classes":["structural","tokenizer","generation-defaults"],"sentence":"Qwen3-32B-AWQ: EOS changed from 151645 to [151645,151643]; anything fine-tuned or served against the old value stops at a different token; BOS declared as 151643 (was unset); context window 65536 to 40960 tokens (2025-05-01).","keys":[{"file":"config","key":"bos_token_id","before":null,"after":151643},{"file":"config","key":"max_position_embeddings","before":65536,"after":40960},{"file":"config","key":"pad_token_id","before":151643,"after":null},{"file":"generation","key":"bos_token_id","before":null,"after":151643},{"file":"generation","key":"eos_token_id","before":151645,"after":[151645,151643]},{"file":"generation","key":"temperature","before":null,"after":0.6},{"file":"generation","key":"top_k","before":null,"after":20},{"file":"generation","key":"top_p","before":null,"after":0.95}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":{"before":65536,"after":40960},"eos":{"before":151645,"after":[151645,151643]},"bos":{"before":null,"after":151643}},"fromSha":"52e74a64c09a6eacff503a0b017ec203767373c4","toSha":"d5f8dfea94a4df504de08bd3670cd5fdfb6a92ce","commit":"https://huggingface.co/Qwen/Qwen3-32B-AWQ/commit/d5f8dfea94a4df504de08bd3670cd5fdfb6a92ce","source":"backfill"},{"id":"Qwen/Qwen3-14B-AWQ","date":"2025-05-01T14:28:28.000Z","classes":["structural","tokenizer","generation-defaults"],"sentence":"Qwen3-14B-AWQ: EOS changed from 151645 to [151645,151643]; anything fine-tuned or served against the old value stops at a different token; BOS declared as 151643 (was unset); context window 65536 to 40960 tokens (2025-05-01).","keys":[{"file":"config","key":"bos_token_id","before":null,"after":151643},{"file":"config","key":"max_position_embeddings","before":65536,"after":40960},{"file":"config","key":"pad_token_id","before":151643,"after":null},{"file":"generation","key":"bos_token_id","before":null,"after":151643},{"file":"generation","key":"eos_token_id","before":151645,"after":[151645,151643]},{"file":"generation","key":"temperature","before":null,"after":0.6},{"file":"generation","key":"top_k","before":null,"after":20},{"file":"generation","key":"top_p","before":null,"after":0.95}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":{"before":65536,"after":40960},"eos":{"before":151645,"after":[151645,151643]},"bos":{"before":null,"after":151643}},"fromSha":"e49110c26e08f396a53e044e12f8093d1502797f","toSha":"bac7b87017e34938f468b4ba40b49247c83fa44a","commit":"https://huggingface.co/Qwen/Qwen3-14B-AWQ/commit/bac7b87017e34938f468b4ba40b49247c83fa44a","source":"backfill"},{"id":"Qwen/Qwen3-8B-Base","date":"2025-04-28T18:19:03.000Z","classes":["template"],"sentence":"Qwen3-8B-Base: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2025-04-28).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"44d5f08f3f72b837","after":"87a2728cb8dc9fe4"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"dfae426db3be706f499449da66cc706d506f4481","toSha":"a64df03b50f8fdeb56e303ed2dd5228dc149f926","commit":"https://huggingface.co/Qwen/Qwen3-8B-Base/commit/a64df03b50f8fdeb56e303ed2dd5228dc149f926","source":"backfill"},{"id":"Qwen/Qwen3-8B-FP8","date":"2025-04-28T18:16:15.000Z","classes":["tokenizer","generation-defaults"],"sentence":"Qwen3-8B-FP8: EOS changed from 151645 to [151645,151643]; anything fine-tuned or served against the old value stops at a different token; BOS declared as 151643 (was unset); generation defaults changed (do_sample unset to true, temperature unset to 0.6, top_k unset to 20, top_p unset to 0.95); a call that does not pass these explicitly now generates differently (2025-04-28).","keys":[{"file":"generation","key":"bos_token_id","before":null,"after":151643},{"file":"generation","key":"do_sample","before":null,"after":true},{"file":"generation","key":"eos_token_id","before":151645,"after":[151645,151643]},{"file":"generation","key":"temperature","before":null,"after":0.6},{"file":"generation","key":"top_k","before":null,"after":20},{"file":"generation","key":"top_p","before":null,"after":0.95}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":{"before":151645,"after":[151645,151643]},"bos":{"before":null,"after":151643}},"fromSha":"18a569c78b184814a5dad8f7f00974458e266cc6","toSha":"add3e8dee80bcfc07aca2fd9edb299ae5c1cb8f7","commit":"https://huggingface.co/Qwen/Qwen3-8B-FP8/commit/add3e8dee80bcfc07aca2fd9edb299ae5c1cb8f7","source":"backfill"},{"id":"Qwen/Qwen3-4B-Base","date":"2025-04-28T18:12:31.000Z","classes":["template"],"sentence":"Qwen3-4B-Base: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2025-04-28).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"44d5f08f3f72b837","after":"87a2728cb8dc9fe4"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"2fcc0a0516f96c6aa5a7677ee212b5c14ec82b77","toSha":"08a151eb7f61cd599444a3fb2fe2008762111300","commit":"https://huggingface.co/Qwen/Qwen3-4B-Base/commit/08a151eb7f61cd599444a3fb2fe2008762111300","source":"backfill"},{"id":"Qwen/Qwen3-1.7B-Base","date":"2025-04-28T17:42:43.000Z","classes":["template"],"sentence":"Qwen3-1.7B-Base: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2025-04-28).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"44d5f08f3f72b837","after":"87a2728cb8dc9fe4"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"658b6ee46470b73f3e0e04e57fc9fb7031fd7daa","toSha":"36be17a0ee54955c2d50eb4af5a126b429874a6e","commit":"https://huggingface.co/Qwen/Qwen3-1.7B-Base/commit/36be17a0ee54955c2d50eb4af5a126b429874a6e","source":"backfill"},{"id":"Qwen/Qwen3-0.6B-Base","date":"2025-04-28T17:30:06.000Z","classes":["template"],"sentence":"Qwen3-0.6B-Base: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2025-04-28).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"44d5f08f3f72b837","after":"87a2728cb8dc9fe4"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"da05741fc6420bf928f7e1eaacd5210b8a4c1b64","toSha":"527f918d05b5c4bcfc25c13b054f06ff8e906e82","commit":"https://huggingface.co/Qwen/Qwen3-0.6B-Base/commit/527f918d05b5c4bcfc25c13b054f06ff8e906e82","source":"backfill"},{"id":"Qwen/Qwen3-235B-A22B","date":"2025-04-28T15:55:17.000Z","classes":["structural"],"sentence":"Qwen3-235B-A22B: MLP width 8192 to 12288 (2025-04-28).","keys":[{"file":"config","key":"intermediate_size","before":8192,"after":12288}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"732a7f4dba756cf2cb75b309d5df0726b9a86990","toSha":"d0aa57eb283a2d2489020470fde653a310b4ec6a","commit":"https://huggingface.co/Qwen/Qwen3-235B-A22B/commit/d0aa57eb283a2d2489020470fde653a310b4ec6a","source":"backfill"},{"id":"Qwen/Qwen3-30B-A3B","date":"2025-04-28T15:50:18.000Z","classes":["structural"],"sentence":"Qwen3-30B-A3B: MLP width 8192 to 6144 (2025-04-28).","keys":[{"file":"config","key":"intermediate_size","before":8192,"after":6144}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"fd4bf3be3236d0145638d8105f2ee424efeda2e3","toSha":"d47d535f78ec44bd57128f8e8aeba17eeb0285ea","commit":"https://huggingface.co/Qwen/Qwen3-30B-A3B/commit/d47d535f78ec44bd57128f8e8aeba17eeb0285ea","source":"backfill"},{"id":"ibm-granite/granite-3.3-2b-instruct","date":"2025-04-24T20:39:36.000Z","classes":["template"],"sentence":"granite-3.3-2b-instruct: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2025-04-24).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"8a58e18acd9ca00a","after":"0bea7711316005a1"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"df305797818a239f983de1d00f49053a0afd7873","toSha":"c0f4eed26da5d7e59889fd66978f1050ed572635","commit":"https://huggingface.co/ibm-granite/granite-3.3-2b-instruct/commit/c0f4eed26da5d7e59889fd66978f1050ed572635","source":"backfill"},{"id":"unsloth/Llama-3.2-3B-Instruct","date":"2025-04-24T19:02:35.000Z","classes":["tokenizer"],"sentence":"Llama-3.2-3B-Instruct: tokenizer_class \"PreTrainedTokenizerFast\" to \"PreTrainedTokenizer\" (2025-04-24).","keys":[{"file":"tokenizer","key":"tokenizer_class","before":"PreTrainedTokenizerFast","after":"PreTrainedTokenizer"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"889fdd323e8975eec1e1e0fc821a2c818e6ad494","toSha":"d577994a547983fdcfad7f2cc5db2d324f8e47dd","commit":"https://huggingface.co/unsloth/Llama-3.2-3B-Instruct/commit/d577994a547983fdcfad7f2cc5db2d324f8e47dd","source":"backfill"},{"id":"moonshotai/Kimi-VL-A3B-Instruct","date":"2025-04-20T12:28:03.000Z","classes":["tokenizer"],"sentence":"Kimi-VL-A3B-Instruct: EOS declared as [163585] (was unset); generation that relied on the library default now stops on this token; BOS declared as 163584 (was unset); pad token unset to 163838; batched training masks a different id (2025-04-20).","keys":[{"file":"generation","key":"bos_token_id","before":null,"after":163584},{"file":"generation","key":"eos_token_id","before":null,"after":[163585]},{"file":"generation","key":"pad_token_id","before":null,"after":163838}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":{"before":null,"after":[163585]},"bos":{"before":null,"after":163584}},"fromSha":"b0b86ff5487fc5b15f6882cb45056918187212aa","toSha":"7a3c132a7b0f1f1677f5a72f258bd3afded7d357","commit":"https://huggingface.co/moonshotai/Kimi-VL-A3B-Instruct/commit/7a3c132a7b0f1f1677f5a72f258bd3afded7d357","source":"backfill"},{"id":"moonshotai/Kimi-VL-A3B-Instruct","date":"2025-04-20T06:37:56.000Z","classes":["generation-defaults"],"sentence":"Kimi-VL-A3B-Instruct: generation defaults changed (temperature 0.6 to 0.2); a call that does not pass these explicitly now generates differently (2025-04-20).","keys":[{"file":"generation","key":"temperature","before":0.6,"after":0.2}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"0157cd86c86ae1151c465f028e6580d99466dba2","toSha":"b0b86ff5487fc5b15f6882cb45056918187212aa","commit":"https://huggingface.co/moonshotai/Kimi-VL-A3B-Instruct/commit/b0b86ff5487fc5b15f6882cb45056918187212aa","source":"backfill"},{"id":"moonshotai/Kimi-VL-A3B-Instruct","date":"2025-04-20T06:31:15.000Z","classes":["generation-defaults"],"sentence":"Kimi-VL-A3B-Instruct: generation defaults changed (do_sample unset to true, temperature 0.2 to 0.6); a call that does not pass these explicitly now generates differently (2025-04-20).","keys":[{"file":"generation","key":"do_sample","before":null,"after":true},{"file":"generation","key":"temperature","before":0.2,"after":0.6}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"b6e11f3a8d61d943e937fbc08551a2fa0c8bee2e","toSha":"0157cd86c86ae1151c465f028e6580d99466dba2","commit":"https://huggingface.co/moonshotai/Kimi-VL-A3B-Instruct/commit/0157cd86c86ae1151c465f028e6580d99466dba2","source":"backfill"},{"id":"ibm-granite/granite-3.3-2b-instruct","date":"2025-04-10T22:23:20.000Z","classes":["template"],"sentence":"granite-3.3-2b-instruct: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2025-04-10).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"6bc46d1fc4c69468","after":"8a58e18acd9ca00a"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"652c333dc5066f2a1764854a1bcd0ce67163d74f","toSha":"39dcd68829433f94f6fe25a51d30825f12985e6d","commit":"https://huggingface.co/ibm-granite/granite-3.3-2b-instruct/commit/39dcd68829433f94f6fe25a51d30825f12985e6d","source":"backfill"},{"id":"ibm-granite/granite-3.3-2b-instruct","date":"2025-04-10T04:02:40.000Z","classes":["template"],"sentence":"granite-3.3-2b-instruct: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2025-04-10).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"14aabd581fb85866","after":"6bc46d1fc4c69468"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"b6554dcbaf706fbda853b88f7a8e940a685d293c","toSha":"652c333dc5066f2a1764854a1bcd0ce67163d74f","commit":"https://huggingface.co/ibm-granite/granite-3.3-2b-instruct/commit/652c333dc5066f2a1764854a1bcd0ce67163d74f","source":"backfill"},{"id":"deepseek-ai/DeepSeek-V3-0324","date":"2025-03-25T06:58:36.000Z","classes":["tokenizer"],"sentence":"DeepSeek-V3-0324: tokenizer model_max_length 16384 to 131072; truncation defaults change (2025-03-25).","keys":[{"file":"tokenizer","key":"model_max_length","before":16384,"after":131072}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"b9763d9f816806a33157f05ea655c818ddce85aa","toSha":"500c51d54664c2d53011337cb84b5b353855bced","commit":"https://huggingface.co/deepseek-ai/DeepSeek-V3-0324/commit/500c51d54664c2d53011337cb84b5b353855bced","source":"backfill"},{"id":"google/gemma-3-1b-it","date":"2025-03-19T15:45:27.000Z","classes":["generation-defaults"],"sentence":"gemma-3-1b-it: generation defaults changed (do_sample unset to true, top_k unset to 64, top_p unset to 0.95); a call that does not pass these explicitly now generates differently (2025-03-19).","keys":[{"file":"generation","key":"do_sample","before":null,"after":true},{"file":"generation","key":"top_k","before":null,"after":64},{"file":"generation","key":"top_p","before":null,"after":0.95}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"9b99be88fdd7a2496bf644baade44348ad736c95","toSha":"417a7e01fafd60e69b3a8e8d868c26960467c11c","commit":"https://huggingface.co/google/gemma-3-1b-it/commit/417a7e01fafd60e69b3a8e8d868c26960467c11c","source":"backfill"},{"id":"google/gemma-3-1b-it","date":"2025-03-12T14:50:25.000Z","classes":["tokenizer"],"sentence":"gemma-3-1b-it: EOS changed from 1 to [1,106]; anything fine-tuned or served against the old value stops at a different token (2025-03-12).","keys":[{"file":"config","key":"eos_token_id","before":1,"after":[1,106]},{"file":"generation","key":"eos_token_id","before":1,"after":[1,106]}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":{"before":1,"after":[1,106]},"bos":null},"fromSha":"ca8476f4df74475f1809d85359c7e50010b6a886","toSha":"9b99be88fdd7a2496bf644baade44348ad736c95","commit":"https://huggingface.co/google/gemma-3-1b-it/commit/9b99be88fdd7a2496bf644baade44348ad736c95","source":"backfill"},{"id":"google/gemma-3-1b-it","date":"2025-03-12T06:31:03.000Z","classes":["template"],"sentence":"gemma-3-1b-it: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2025-03-12).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"5437d0ad30bb8db6","after":"7de1c58e208eda46"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"6c0ce4ee3210e2b4d8f7b153b3111efa472c58b5","toSha":"ca8476f4df74475f1809d85359c7e50010b6a886","commit":"https://huggingface.co/google/gemma-3-1b-it/commit/ca8476f4df74475f1809d85359c7e50010b6a886","source":"backfill"},{"id":"unsloth/Llama-3.2-1B-Instruct","date":"2025-02-25T02:04:27.000Z","classes":["tokenizer"],"sentence":"Llama-3.2-1B-Instruct: tokenizer_class \"PreTrainedTokenizerFast\" to \"PreTrainedTokenizer\" (2025-02-25).","keys":[{"file":"tokenizer","key":"tokenizer_class","before":"PreTrainedTokenizerFast","after":"PreTrainedTokenizer"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"eb49081324edb2ff14f848ce16393c067c6f4976","toSha":"9b58d4a36161a1e49ecf0a69d20b2736fef8e438","commit":"https://huggingface.co/unsloth/Llama-3.2-1B-Instruct/commit/9b58d4a36161a1e49ecf0a69d20b2736fef8e438","source":"backfill"},{"id":"google/siglip2-giant-opt-patch16-384","date":"2025-02-18T10:36:56.000Z","classes":["structural"],"sentence":"siglip2-giant-opt-patch16-384: architectures [\"SiglipModel\"] to unset; a different class is instantiated from the same id; declared dtype \"float32\" to unset (2025-02-18).","keys":[{"file":"config","key":"architectures","before":["SiglipModel"],"after":null},{"file":"config","key":"dtype","before":"float32","after":null}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"c6cbb7de94d2c4213fc916cfebc6e1deb4a5cb3b","toSha":"f3bbb46b2502dadec52d98c4e59724748623d470","commit":"https://huggingface.co/google/siglip2-giant-opt-patch16-384/commit/f3bbb46b2502dadec52d98c4e59724748623d470","source":"backfill"},{"id":"unsloth/Meta-Llama-3.1-8B-Instruct","date":"2025-02-15T10:24:46.000Z","classes":["tokenizer"],"sentence":"Meta-Llama-3.1-8B-Instruct: tokenizer_class \"PreTrainedTokenizerFast\" to \"PreTrainedTokenizer\" (2025-02-15).","keys":[{"file":"tokenizer","key":"tokenizer_class","before":"PreTrainedTokenizerFast","after":"PreTrainedTokenizer"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"28e14b5ea9119456491adc638b85594d58532588","toSha":"a2856192dd7c25b842431f39c179a6c2c2f627d1","commit":"https://huggingface.co/unsloth/Meta-Llama-3.1-8B-Instruct/commit/a2856192dd7c25b842431f39c179a6c2c2f627d1","source":"backfill"},{"id":"deepseek-ai/DeepSeek-R1-Distill-Qwen-7B","date":"2025-02-09T06:34:39.000Z","classes":["template"],"sentence":"DeepSeek-R1-Distill-Qwen-7B: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2025-02-09).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"b6835114b7303ddd","after":"56a1447ad31926fd"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"4831ee1375be5b4ff5a4abf7984e13628db44e35","toSha":"6602cadec947dbb53e64f3d8d6425320b2197247","commit":"https://huggingface.co/deepseek-ai/DeepSeek-R1-Distill-Qwen-7B/commit/6602cadec947dbb53e64f3d8d6425320b2197247","source":"backfill"},{"id":"deepseek-ai/DeepSeek-R1-Distill-Qwen-32B","date":"2025-02-09T06:34:35.000Z","classes":["template"],"sentence":"DeepSeek-R1-Distill-Qwen-32B: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2025-02-09).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"b6835114b7303ddd","after":"56a1447ad31926fd"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"ca24ee48c0532a014ddea4c32437a6f4be981ab2","toSha":"3865e12a1eb7cbd641ab3f9dfc28c588c6b0c1e9","commit":"https://huggingface.co/deepseek-ai/DeepSeek-R1-Distill-Qwen-32B/commit/3865e12a1eb7cbd641ab3f9dfc28c588c6b0c1e9","source":"backfill"},{"id":"deepseek-ai/DeepSeek-R1-Distill-Qwen-14B","date":"2025-02-09T06:34:31.000Z","classes":["template"],"sentence":"DeepSeek-R1-Distill-Qwen-14B: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2025-02-09).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"b6835114b7303ddd","after":"56a1447ad31926fd"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"f37b6a67b659a9445035649e6064005c312543ae","toSha":"5ee96d8a09692e87087a6e0496d87124a1cdc3fe","commit":"https://huggingface.co/deepseek-ai/DeepSeek-R1-Distill-Qwen-14B/commit/5ee96d8a09692e87087a6e0496d87124a1cdc3fe","source":"backfill"},{"id":"deepseek-ai/DeepSeek-R1-Distill-Qwen-1.5B","date":"2025-02-09T06:34:28.000Z","classes":["template"],"sentence":"DeepSeek-R1-Distill-Qwen-1.5B: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2025-02-09).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"b6835114b7303ddd","after":"56a1447ad31926fd"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"164af3f02f5b3aedf8cbdb94aee9e2705ae31be5","toSha":"530ca3e1ad39d440e182c2e4317aa40f012512fa","commit":"https://huggingface.co/deepseek-ai/DeepSeek-R1-Distill-Qwen-1.5B/commit/530ca3e1ad39d440e182c2e4317aa40f012512fa","source":"backfill"},{"id":"deepseek-ai/DeepSeek-R1","date":"2025-02-09T06:34:11.000Z","classes":["template"],"sentence":"DeepSeek-R1: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2025-02-09).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"3a3036eeb96ca0e4","after":"563b1ce7d61d50a9"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"cb48aa8cb28c160ec8d853707278e0402c9ad01a","toSha":"8a58a132790c9935686eb97f042afa8013451c9f","commit":"https://huggingface.co/deepseek-ai/DeepSeek-R1/commit/8a58a132790c9935686eb97f042afa8013451c9f","source":"backfill"},{"id":"fixie-ai/ultravox-v0_5-llama-3_2-1b","date":"2025-02-07T00:24:03.000Z","classes":["structural"],"sentence":"ultravox-v0_5-llama-3_2-1b: layer count unset to 16; derived parameters 2.42B to 2.84B (+419.5M, +0.84 GB of bf16 weights) (2025-02-07).","keys":[{"file":"config","key":"num_hidden_layers","before":null,"after":16}],"consequence":{"params":{"before":2418734336,"after":2838197504},"weightsGbBf16":{"before":4.837,"after":5.676},"trainPeakGb":{"before":79.553,"after":96.445},"fingerprint":{"before":"1c881eaf","after":"eefd22cc"},"context":null,"eos":null,"bos":null},"fromSha":"f27f705716a3aebde26fed3e3f1966e99f4a59c6","toSha":"0ace0e297729be51dda27bfb2fd26442d35e807c","commit":"https://huggingface.co/fixie-ai/ultravox-v0_5-llama-3_2-1b/commit/0ace0e297729be51dda27bfb2fd26442d35e807c","source":"backfill"},{"id":"fixie-ai/ultravox-v0_5-llama-3_2-1b","date":"2025-02-07T00:23:10.000Z","classes":["structural"],"sentence":"ultravox-v0_5-llama-3_2-1b: attention heads unset to 32; derived parameters 2.42B to 2.42B (+786.4K, +0.00 GB of bf16 weights) (2025-02-07).","keys":[{"file":"config","key":"num_attention_heads","before":null,"after":32}],"consequence":{"params":{"before":2417947904,"after":2418734336},"weightsGbBf16":{"before":4.836,"after":4.837},"trainPeakGb":{"before":79.525,"after":79.553},"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"a5eeb6d7918b59338df8dcefb9a95d7e0d16193a","toSha":"f27f705716a3aebde26fed3e3f1966e99f4a59c6","commit":"https://huggingface.co/fixie-ai/ultravox-v0_5-llama-3_2-1b/commit/f27f705716a3aebde26fed3e3f1966e99f4a59c6","source":"backfill"},{"id":"microsoft/phi-4","date":"2025-02-04T11:51:00.000Z","classes":["tokenizer","template"],"sentence":"phi-4: EOS changed from \"<|endoftext|>\" to \"<|im_end|>\"; anything fine-tuned or served against the old value stops at a different token; pad token \"<|endoftext|>\" to \"<|dummy_85|>\"; batched training masks a different id; the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2025-02-04).","keys":[{"file":"config","key":"eos_token_id","before":100257,"after":100265},{"file":"config","key":"pad_token_id","before":100257,"after":100349},{"file":"tokenizer","key":"chat_template_sha256","before":"a41fb8727c4f1e72","after":"fe4120415a3858b8"},{"file":"tokenizer","key":"eos_token","before":"<|endoftext|>","after":"<|im_end|>"},{"file":"tokenizer","key":"pad_token","before":"<|endoftext|>","after":"<|dummy_85|>"},{"file":"generation","key":"eos_token_id","before":[100257,100265],"after":100265},{"file":"generation","key":"pad_token_id","before":null,"after":100349}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":{"before":"<|endoftext|>","after":"<|im_end|>"},"bos":null},"fromSha":"381727a5ee103da6c1b14ecd3d39cd09832cbcf8","toSha":"6fbb3d3bbe726c99b4188087b4deeec1bceac5ae","commit":"https://huggingface.co/microsoft/phi-4/commit/6fbb3d3bbe726c99b4188087b4deeec1bceac5ae","source":"backfill"},{"id":"unsloth/Meta-Llama-3.1-8B-Instruct","date":"2025-02-02T04:35:10.000Z","classes":["structural","tokenizer"],"sentence":"Meta-Llama-3.1-8B-Instruct: EOS changed from [128001,128008,128009] to 128009; anything fine-tuned or served against the old value stops at a different token; head dim unset to 128; add_bos_token unset to true; encoding the same text yields different ids (2025-02-02).","keys":[{"file":"config","key":"eos_token_id","before":[128001,128008,128009],"after":128009},{"file":"config","key":"head_dim","before":null,"after":128},{"file":"tokenizer","key":"add_bos_token","before":null,"after":true}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":{"before":[128001,128008,128009],"after":128009},"bos":null},"fromSha":"62b4abd08ef28d337b64d3d668f040b75fb6fadc","toSha":"28e14b5ea9119456491adc638b85594d58532588","commit":"https://huggingface.co/unsloth/Meta-Llama-3.1-8B-Instruct/commit/28e14b5ea9119456491adc638b85594d58532588","source":"backfill"},{"id":"unsloth/Llama-3.2-3B-Instruct","date":"2025-01-23T05:18:10.000Z","classes":["tokenizer"],"sentence":"Llama-3.2-3B-Instruct: EOS changed from [128001,128008,128009] to 128009; anything fine-tuned or served against the old value stops at a different token; add_bos_token unset to true; encoding the same text yields different ids (2025-01-23).","keys":[{"file":"config","key":"eos_token_id","before":[128001,128008,128009],"after":128009},{"file":"tokenizer","key":"add_bos_token","before":null,"after":true}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":{"before":[128001,128008,128009],"after":128009},"bos":null},"fromSha":"13da286fb6e4d39a98ff3dab59372531b7c7e566","toSha":"889fdd323e8975eec1e1e0fc821a2c818e6ad494","commit":"https://huggingface.co/unsloth/Llama-3.2-3B-Instruct/commit/889fdd323e8975eec1e1e0fc821a2c818e6ad494","source":"backfill"},{"id":"unsloth/Llama-3.2-1B-Instruct","date":"2025-01-23T04:59:34.000Z","classes":["tokenizer"],"sentence":"Llama-3.2-1B-Instruct: EOS changed from [128001,128008,128009] to 128009; anything fine-tuned or served against the old value stops at a different token (2025-01-23).","keys":[{"file":"config","key":"eos_token_id","before":[128001,128008,128009],"after":128009}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":{"before":[128001,128008,128009],"after":128009},"bos":null},"fromSha":"2906abdacc8c773de918615c4c5f7232f1109e62","toSha":"eb49081324edb2ff14f848ce16393c067c6f4976","commit":"https://huggingface.co/unsloth/Llama-3.2-1B-Instruct/commit/eb49081324edb2ff14f848ce16393c067c6f4976","source":"backfill"},{"id":"unsloth/Llama-3.2-1B-Instruct","date":"2025-01-22T12:11:26.000Z","classes":["template"],"sentence":"Llama-3.2-1B-Instruct: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2025-01-22).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"053402b579e589c0","after":"5816fce10444e03c"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"0430e7102ae7f4fad2b056a4d5064abd985588fc","toSha":"2906abdacc8c773de918615c4c5f7232f1109e62","commit":"https://huggingface.co/unsloth/Llama-3.2-1B-Instruct/commit/2906abdacc8c773de918615c4c5f7232f1109e62","source":"backfill"},{"id":"deepseek-ai/DeepSeek-R1-Distill-Qwen-7B","date":"2025-01-22T11:22:19.000Z","classes":["tokenizer"],"sentence":"DeepSeek-R1-Distill-Qwen-7B: add_bos_token false to true; encoding the same text yields different ids (2025-01-22).","keys":[{"file":"tokenizer","key":"add_bos_token","before":false,"after":true}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"b640c47fb9acbc6333ced6216252a17ed3abc4f6","toSha":"4831ee1375be5b4ff5a4abf7984e13628db44e35","commit":"https://huggingface.co/deepseek-ai/DeepSeek-R1-Distill-Qwen-7B/commit/4831ee1375be5b4ff5a4abf7984e13628db44e35","source":"backfill"},{"id":"deepseek-ai/DeepSeek-R1-Distill-Qwen-32B","date":"2025-01-22T11:22:17.000Z","classes":["tokenizer"],"sentence":"DeepSeek-R1-Distill-Qwen-32B: add_bos_token false to true; encoding the same text yields different ids (2025-01-22).","keys":[{"file":"tokenizer","key":"add_bos_token","before":false,"after":true}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"d5e5623342de724261521aac455648c987666362","toSha":"ca24ee48c0532a014ddea4c32437a6f4be981ab2","commit":"https://huggingface.co/deepseek-ai/DeepSeek-R1-Distill-Qwen-32B/commit/ca24ee48c0532a014ddea4c32437a6f4be981ab2","source":"backfill"},{"id":"deepseek-ai/DeepSeek-R1-Distill-Qwen-14B","date":"2025-01-22T11:22:14.000Z","classes":["tokenizer"],"sentence":"DeepSeek-R1-Distill-Qwen-14B: add_bos_token false to true; encoding the same text yields different ids (2025-01-22).","keys":[{"file":"tokenizer","key":"add_bos_token","before":false,"after":true}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"be829b6f685e5d08ba1a078ba6f2d7e3a3b61478","toSha":"f37b6a67b659a9445035649e6064005c312543ae","commit":"https://huggingface.co/deepseek-ai/DeepSeek-R1-Distill-Qwen-14B/commit/f37b6a67b659a9445035649e6064005c312543ae","source":"backfill"},{"id":"deepseek-ai/DeepSeek-R1-Distill-Qwen-1.5B","date":"2025-01-22T11:22:12.000Z","classes":["tokenizer"],"sentence":"DeepSeek-R1-Distill-Qwen-1.5B: add_bos_token false to true; encoding the same text yields different ids (2025-01-22).","keys":[{"file":"tokenizer","key":"add_bos_token","before":false,"after":true}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"e93784059177e379b713658b88145e0028fac5c1","toSha":"164af3f02f5b3aedf8cbdb94aee9e2705ae31be5","commit":"https://huggingface.co/deepseek-ai/DeepSeek-R1-Distill-Qwen-1.5B/commit/164af3f02f5b3aedf8cbdb94aee9e2705ae31be5","source":"backfill"},{"id":"deepseek-ai/DeepSeek-R1","date":"2025-01-22T11:21:42.000Z","classes":["tokenizer"],"sentence":"DeepSeek-R1: add_bos_token false to true; encoding the same text yields different ids (2025-01-22).","keys":[{"file":"tokenizer","key":"add_bos_token","before":false,"after":true}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"7bae4bacd0f64d020364cc43d9560588893e94cc","toSha":"cb48aa8cb28c160ec8d853707278e0402c9ad01a","commit":"https://huggingface.co/deepseek-ai/DeepSeek-R1/commit/cb48aa8cb28c160ec8d853707278e0402c9ad01a","source":"backfill"},{"id":"Alibaba-NLP/gte-reranker-modernbert-base","date":"2025-01-22T10:44:55.000Z","classes":["tokenizer"],"sentence":"gte-reranker-modernbert-base: tokenizer model_max_length 1e+30 to 8192; truncation defaults change (2025-01-22).","keys":[{"file":"tokenizer","key":"model_max_length","before":1e+30,"after":8192}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"c40156962ee2a34679b0c8399e0d1bb9d68d54ab","toSha":"3c50625ccab323fa00faa83ace927e613c690a8f","commit":"https://huggingface.co/Alibaba-NLP/gte-reranker-modernbert-base/commit/3c50625ccab323fa00faa83ace927e613c690a8f","source":"backfill"},{"id":"unsloth/Llama-3.2-1B-Instruct","date":"2025-01-22T09:38:08.000Z","classes":["tokenizer","template","generation-defaults"],"sentence":"Llama-3.2-1B-Instruct: pad token unset to 128004; batched training masks a different id; the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering; add_bos_token unset to true; encoding the same text yields different ids (2025-01-22).","keys":[{"file":"config","key":"pad_token_id","before":null,"after":128004},{"file":"tokenizer","key":"add_bos_token","before":null,"after":true},{"file":"tokenizer","key":"chat_template_sha256","before":"5816fce10444e03c","after":"053402b579e589c0"},{"file":"generation","key":"max_length","before":null,"after":131072},{"file":"generation","key":"pad_token_id","before":null,"after":128004}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"bd6d97814a86488154b2032080b776ca831e5547","toSha":"0430e7102ae7f4fad2b056a4d5064abd985588fc","commit":"https://huggingface.co/unsloth/Llama-3.2-1B-Instruct/commit/0430e7102ae7f4fad2b056a4d5064abd985588fc","source":"backfill"},{"id":"deepseek-ai/DeepSeek-R1-Distill-Qwen-7B","date":"2025-01-22T02:36:29.000Z","classes":["tokenizer"],"sentence":"DeepSeek-R1-Distill-Qwen-7B: add_bos_token true to false; encoding the same text yields different ids (2025-01-22).","keys":[{"file":"tokenizer","key":"add_bos_token","before":true,"after":false}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"4e5485ed9bc7191a570b78c3bf4b9b252ca63793","toSha":"b640c47fb9acbc6333ced6216252a17ed3abc4f6","commit":"https://huggingface.co/deepseek-ai/DeepSeek-R1-Distill-Qwen-7B/commit/b640c47fb9acbc6333ced6216252a17ed3abc4f6","source":"backfill"},{"id":"deepseek-ai/DeepSeek-R1-Distill-Qwen-32B","date":"2025-01-22T02:36:27.000Z","classes":["tokenizer"],"sentence":"DeepSeek-R1-Distill-Qwen-32B: add_bos_token true to false; encoding the same text yields different ids (2025-01-22).","keys":[{"file":"tokenizer","key":"add_bos_token","before":true,"after":false}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"2d78713b01ecefe27a89fafec248a5dfd731396f","toSha":"d5e5623342de724261521aac455648c987666362","commit":"https://huggingface.co/deepseek-ai/DeepSeek-R1-Distill-Qwen-32B/commit/d5e5623342de724261521aac455648c987666362","source":"backfill"},{"id":"deepseek-ai/DeepSeek-R1-Distill-Qwen-14B","date":"2025-01-22T02:36:25.000Z","classes":["tokenizer"],"sentence":"DeepSeek-R1-Distill-Qwen-14B: add_bos_token true to false; encoding the same text yields different ids (2025-01-22).","keys":[{"file":"tokenizer","key":"add_bos_token","before":true,"after":false}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"5bbe9fe8da98def5e672fb9c295d5303bb6ff5ec","toSha":"be829b6f685e5d08ba1a078ba6f2d7e3a3b61478","commit":"https://huggingface.co/deepseek-ai/DeepSeek-R1-Distill-Qwen-14B/commit/be829b6f685e5d08ba1a078ba6f2d7e3a3b61478","source":"backfill"},{"id":"deepseek-ai/DeepSeek-R1-Distill-Qwen-1.5B","date":"2025-01-22T02:36:23.000Z","classes":["tokenizer"],"sentence":"DeepSeek-R1-Distill-Qwen-1.5B: add_bos_token true to false; encoding the same text yields different ids (2025-01-22).","keys":[{"file":"tokenizer","key":"add_bos_token","before":true,"after":false}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"73d63fa6f97d4285b49bdbedf6a161ceba1e55ad","toSha":"e93784059177e379b713658b88145e0028fac5c1","commit":"https://huggingface.co/deepseek-ai/DeepSeek-R1-Distill-Qwen-1.5B/commit/e93784059177e379b713658b88145e0028fac5c1","source":"backfill"},{"id":"deepseek-ai/DeepSeek-R1","date":"2025-01-22T02:36:01.000Z","classes":["tokenizer"],"sentence":"DeepSeek-R1: add_bos_token true to false; encoding the same text yields different ids (2025-01-22).","keys":[{"file":"tokenizer","key":"add_bos_token","before":true,"after":false}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"ebc477ae6dc22c217b3b3990effcab0112a67273","toSha":"7bae4bacd0f64d020364cc43d9560588893e94cc","commit":"https://huggingface.co/deepseek-ai/DeepSeek-R1/commit/7bae4bacd0f64d020364cc43d9560588893e94cc","source":"backfill"},{"id":"unsloth/Llama-3.2-3B-Instruct","date":"2025-01-16T09:19:20.000Z","classes":["tokenizer","generation-defaults"],"sentence":"Llama-3.2-3B-Instruct: pad token unset to 128004; batched training masks a different id; generation defaults changed (max_length unset to 131072); a call that does not pass these explicitly now generates differently (2025-01-16).","keys":[{"file":"config","key":"pad_token_id","before":null,"after":128004},{"file":"generation","key":"max_length","before":null,"after":131072},{"file":"generation","key":"pad_token_id","before":null,"after":128004}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"cfe38f8f38afea8b9f9cc66f634bf53c321cf426","toSha":"c64b6bcd612f3348d4ae448944518845e6478eeb","commit":"https://huggingface.co/unsloth/Llama-3.2-3B-Instruct/commit/c64b6bcd612f3348d4ae448944518845e6478eeb","source":"backfill"},{"id":"answerdotai/ModernBERT-base","date":"2025-01-15T20:11:48.000Z","classes":["tokenizer"],"sentence":"ModernBERT-base: tokenizer model_max_length 1e+30 to 8192; truncation defaults change (2025-01-15).","keys":[{"file":"tokenizer","key":"model_max_length","before":1e+30,"after":8192}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"94032bb66234a691cf6248265170006a7ced4970","toSha":"8949b909ec900327062f0ebf497f51aef5e6f0c8","commit":"https://huggingface.co/answerdotai/ModernBERT-base/commit/8949b909ec900327062f0ebf497f51aef5e6f0c8","source":"backfill"},{"id":"Alibaba-NLP/gte-Qwen2-7B-instruct","date":"2025-01-11T07:10:38.000Z","classes":["structural"],"sentence":"gte-Qwen2-7B-instruct: config.json was removed; whatever defaults it set now come from the library (2025-01-11).","keys":[{"file":"config","key":"(file)","before":"present","after":null}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"03d916b0194169a2580541896bbfcdf3914ae67b","toSha":"4a5d32e44c4023e09093d0e9b2b4e6eea1f6ad0a","commit":"https://huggingface.co/Alibaba-NLP/gte-Qwen2-7B-instruct/commit/4a5d32e44c4023e09093d0e9b2b4e6eea1f6ad0a","source":"backfill"},{"id":"HuggingFaceTB/SmolLM2-135M-Instruct","date":"2025-01-07T16:27:22.000Z","classes":["tokenizer"],"sentence":"SmolLM2-135M-Instruct: tokenizer model_max_length 2048 to 8192; truncation defaults change (2025-01-07).","keys":[{"file":"tokenizer","key":"model_max_length","before":2048,"after":8192}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"75fd0ae5b521241aac18793eb0d6cb3598d86055","toSha":"ec84f302454dcb013b57c8acd7ca9cb331981e7e","commit":"https://huggingface.co/HuggingFaceTB/SmolLM2-135M-Instruct/commit/ec84f302454dcb013b57c8acd7ca9cb331981e7e","source":"backfill"},{"id":"HuggingFaceTB/SmolLM2-360M-Instruct","date":"2025-01-07T16:27:01.000Z","classes":["tokenizer"],"sentence":"SmolLM2-360M-Instruct: tokenizer model_max_length 2048 to 8192; truncation defaults change (2025-01-07).","keys":[{"file":"tokenizer","key":"model_max_length","before":2048,"after":8192}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"4873f67095301d304753fae05bc09ec766634e50","toSha":"8ab9c9e783a7b6a1a6121310b86bef9b3a12ce9f","commit":"https://huggingface.co/HuggingFaceTB/SmolLM2-360M-Instruct/commit/8ab9c9e783a7b6a1a6121310b86bef9b3a12ce9f","source":"backfill"},{"id":"HuggingFaceTB/SmolLM2-1.7B-Instruct","date":"2025-01-07T16:26:43.000Z","classes":["tokenizer"],"sentence":"SmolLM2-1.7B-Instruct: tokenizer model_max_length 2048 to 8192; truncation defaults change (2025-01-07).","keys":[{"file":"tokenizer","key":"model_max_length","before":2048,"after":8192}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"d1bb90bcfbe0f211109880f4da18da66f229c4f6","toSha":"450ff1f12781f49cb595c70d9e402812c099e748","commit":"https://huggingface.co/HuggingFaceTB/SmolLM2-1.7B-Instruct/commit/450ff1f12781f49cb595c70d9e402812c099e748","source":"backfill"},{"id":"HuggingFaceTB/SmolLM2-135M-Instruct","date":"2025-01-07T16:15:02.000Z","classes":["tokenizer"],"sentence":"SmolLM2-135M-Instruct: tokenizer model_max_length 8192 to 2048; truncation defaults change (2025-01-07).","keys":[{"file":"tokenizer","key":"model_max_length","before":8192,"after":2048}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"9a66ea6ac76f436080ad0be87c98f751bbbc23a3","toSha":"75fd0ae5b521241aac18793eb0d6cb3598d86055","commit":"https://huggingface.co/HuggingFaceTB/SmolLM2-135M-Instruct/commit/75fd0ae5b521241aac18793eb0d6cb3598d86055","source":"backfill"},{"id":"HuggingFaceTB/SmolLM2-135M-Instruct","date":"2025-01-07T16:14:49.000Z","classes":["tokenizer"],"sentence":"SmolLM2-135M-Instruct: tokenizer model_max_length 2048 to 8192; truncation defaults change (2025-01-07).","keys":[{"file":"tokenizer","key":"model_max_length","before":2048,"after":8192}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"5a33ba103645800d7b3790c4448546c1b73efc71","toSha":"9a66ea6ac76f436080ad0be87c98f751bbbc23a3","commit":"https://huggingface.co/HuggingFaceTB/SmolLM2-135M-Instruct/commit/9a66ea6ac76f436080ad0be87c98f751bbbc23a3","source":"backfill"},{"id":"answerdotai/ModernBERT-base","date":"2024-12-19T21:14:57.000Z","classes":["structural"],"sentence":"ModernBERT-base: derived parameters 130.1M to 136.0M (+5.9M, +0.01 GB of bf16 weights); context window 512 to 8192 tokens (2024-12-19).","keys":[{"file":"config","key":"max_position_embeddings","before":512,"after":8192}],"consequence":{"params":{"before":130052352,"after":135950592},"weightsGbBf16":{"before":0.26,"after":0.272},"trainPeakGb":{"before":8.257,"after":25.649},"fingerprint":null,"context":{"before":512,"after":8192},"eos":null,"bos":null},"fromSha":"fa3113d09d6dacfe5528708803dd92a62240237a","toSha":"5756c58a31a2478f9e62146021f48295a92c3da5","commit":"https://huggingface.co/answerdotai/ModernBERT-base/commit/5756c58a31a2478f9e62146021f48295a92c3da5","source":"backfill"},{"id":"ai-sage/Giga-Embeddings-instruct","date":"2024-12-19T13:40:27.000Z","classes":["tokenizer"],"sentence":"Giga-Embeddings-instruct: pad token unset to \"<unk>\"; batched training masks a different id (2024-12-19).","keys":[{"file":"tokenizer","key":"pad_token","before":null,"after":"<unk>"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"b2c9c67d18d6fdf4c387cb1e07f7650214338c2f","toSha":"201baa9d6f8f22c095d09b61fb50228d9382306f","commit":"https://huggingface.co/ai-sage/Giga-Embeddings-instruct/commit/201baa9d6f8f22c095d09b61fb50228d9382306f","source":"backfill"},{"id":"utter-project/EuroLLM-1.7B-Instruct","date":"2024-12-16T12:46:04.000Z","classes":["tokenizer"],"sentence":"EuroLLM-1.7B-Instruct: EOS changed from 2 to 4; anything fine-tuned or served against the old value stops at a different token; BOS changed from 1 to 3; prompts built by the old tokenizer start with a different token (2024-12-16).","keys":[{"file":"generation","key":"bos_token_id","before":1,"after":3},{"file":"generation","key":"eos_token_id","before":2,"after":4}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":{"before":2,"after":4},"bos":{"before":1,"after":3}},"fromSha":"cd9ed3768c25b5f71ea70fc3a6f5e7faab349ed8","toSha":"a25c7fa65fc2a644e6270b8940dbe295b51da681","commit":"https://huggingface.co/utter-project/EuroLLM-1.7B-Instruct/commit/a25c7fa65fc2a644e6270b8940dbe295b51da681","source":"backfill"},{"id":"answerdotai/ModernBERT-base","date":"2024-12-11T15:48:49.000Z","classes":["structural"],"sentence":"ModernBERT-base: architectures [\"ModernBertModel\"] to [\"ModernBertForMaskedLM\"]; a different class is instantiated from the same id (2024-12-11).","keys":[{"file":"config","key":"architectures","before":["ModernBertModel"],"after":["ModernBertForMaskedLM"]}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"cb3183fe4f51d9d94a52e4e36496ff734eac7fa2","toSha":"290243f4fe9b0aa78cc87a27d23351b22bc63c41","commit":"https://huggingface.co/answerdotai/ModernBERT-base/commit/290243f4fe9b0aa78cc87a27d23351b22bc63c41","source":"backfill"},{"id":"microsoft/Florence-2-large","date":"2024-12-08T22:30:48.000Z","classes":["structural"],"sentence":"Florence-2-large: derived parameters 204.7M to 207.8M (+3.1M, +0.01 GB of bf16 weights); context window 1024 to 4096 tokens (2024-12-08).","keys":[{"file":"config","key":"max_position_embeddings","before":1024,"after":4096}],"consequence":{"params":{"before":204698624,"after":207844352},"weightsGbBf16":{"before":0.409,"after":0.416},"trainPeakGb":{"before":17.819,"after":31.768},"fingerprint":null,"context":{"before":1024,"after":4096},"eos":null,"bos":null},"fromSha":"6bf179230dd8855083a51a5e11beb04aec1291fd","toSha":"00d2f1570b00c6dea5df998f5635db96840436bc","commit":"https://huggingface.co/microsoft/Florence-2-large/commit/00d2f1570b00c6dea5df998f5635db96840436bc","source":"backfill"},{"id":"meta-llama/Llama-3.1-405B-FP8","date":"2024-12-06T19:41:17.000Z","classes":["tokenizer"],"sentence":"Llama-3.1-405B-FP8: pad token unset to \"<|finetune_right_pad_id|>\"; batched training masks a different id (2024-12-06).","keys":[{"file":"tokenizer","key":"pad_token","before":null,"after":"<|finetune_right_pad_id|>"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"d79e67b5ee26735510cff4fb9ec10fb74fadb8c1","toSha":"cb4739810a51170037906b5daa2bb78ed13cd43c","commit":"https://huggingface.co/meta-llama/Llama-3.1-405B-FP8/commit/cb4739810a51170037906b5daa2bb78ed13cd43c","source":"backfill"},{"id":"OpenGVLab/InternVL2-1B","date":"2024-12-06T14:33:27.000Z","classes":["structural"],"sentence":"InternVL2-1B: derived parameters 798.6M to 934.6M (+136.0M, +0.27 GB of bf16 weights); tie_word_embeddings true to false; the output head is no longer shared with the embedding (2024-12-06).","keys":[{"file":"config","key":"tie_word_embeddings","before":true,"after":false}],"consequence":{"params":{"before":798583808,"after":934619239},"weightsGbBf16":{"before":1.597,"after":1.869},"trainPeakGb":{"before":159.748,"after":248.672},"fingerprint":{"before":"54aab4e0","after":"023b735d"},"context":null,"eos":null,"bos":null},"fromSha":"8d827904f96b944be1d03fccd82e51a000c10d1d","toSha":"b5e11b0270f697b3dcb33a21ee641bd49ff4b042","commit":"https://huggingface.co/OpenGVLab/InternVL2-1B/commit/b5e11b0270f697b3dcb33a21ee641bd49ff4b042","source":"backfill"},{"id":"OpenGVLab/InternVL2_5-4B","date":"2024-11-29T09:01:26.000Z","classes":["structural"],"sentence":"InternVL2_5-4B: derived parameters 3.39B to 3.70B (+310.8M, +0.62 GB of bf16 weights); tie_word_embeddings true to false; the output head is no longer shared with the embedding (2024-11-29).","keys":[{"file":"config","key":"tie_word_embeddings","before":true,"after":false}],"consequence":{"params":{"before":3391320064,"after":3702102138},"weightsGbBf16":{"before":6.783,"after":7.404},"trainPeakGb":{"before":488.983,"after":580.989},"fingerprint":{"before":"2795de95","after":"ee16fbce"},"context":null,"eos":null,"bos":null},"fromSha":"74b297953c78a25b652f1126267a82b2b3b147ec","toSha":"f01b74c11c5317424c85e41a9b1d06a940e80c4d","commit":"https://huggingface.co/OpenGVLab/InternVL2_5-4B/commit/f01b74c11c5317424c85e41a9b1d06a940e80c4d","source":"backfill"},{"id":"OpenGVLab/InternVL2_5-4B","date":"2024-11-26T10:41:21.000Z","classes":["tokenizer"],"sentence":"InternVL2_5-4B: EOS changed from [151644,151645] to [151644,151645,151643]; anything fine-tuned or served against the old value stops at a different token (2024-11-26).","keys":[{"file":"generation","key":"eos_token_id","before":[151644,151645],"after":[151644,151645,151643]}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":{"before":[151644,151645],"after":[151644,151645,151643]},"bos":null},"fromSha":"e21cfb8cc8760677ca17aa43a90657810d6c53be","toSha":"74b297953c78a25b652f1126267a82b2b3b147ec","commit":"https://huggingface.co/OpenGVLab/InternVL2_5-4B/commit/74b297953c78a25b652f1126267a82b2b3b147ec","source":"backfill"},{"id":"jinaai/jina-clip-v2","date":"2024-11-22T15:25:09.000Z","classes":["structural"],"sentence":"jina-clip-v2: declared dtype \"float16\" to \"bfloat16\" (2024-11-22).","keys":[{"file":"config","key":"dtype","before":"float16","after":"bfloat16"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"8ea414a1965e5e4037205ac7aeae92143d4e970b","toSha":"ced12ef41313ab7adf256b1649abd90b4d49b713","commit":"https://huggingface.co/jinaai/jina-clip-v2/commit/ced12ef41313ab7adf256b1649abd90b4d49b713","source":"backfill"},{"id":"jinaai/jina-clip-v2","date":"2024-11-22T14:59:43.000Z","classes":["structural"],"sentence":"jina-clip-v2: declared dtype \"bfloat16\" to \"float16\" (2024-11-22).","keys":[{"file":"config","key":"dtype","before":"bfloat16","after":"float16"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"87eae09148285716b700004506fda9dd3a479cdb","toSha":"8ea414a1965e5e4037205ac7aeae92143d4e970b","commit":"https://huggingface.co/jinaai/jina-clip-v2/commit/8ea414a1965e5e4037205ac7aeae92143d4e970b","source":"backfill"},{"id":"jinaai/jina-clip-v2","date":"2024-11-22T14:54:42.000Z","classes":["metadata"],"sentence":"jina-clip-v2: declared dtype unset to \"bfloat16\" (2024-11-22).","keys":[{"file":"config","key":"dtype","before":null,"after":"bfloat16"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"3f4007abf78ee1c2be280e6c36fc9064d1fd6d11","toSha":"87eae09148285716b700004506fda9dd3a479cdb","commit":"https://huggingface.co/jinaai/jina-clip-v2/commit/87eae09148285716b700004506fda9dd3a479cdb","source":"backfill"},{"id":"Snowflake/snowflake-arctic-embed-m-v2.0","date":"2024-11-18T19:20:39.000Z","classes":["structural"],"sentence":"snowflake-arctic-embed-m-v2.0: architectures [\"NewModel\"] to [\"GteModel\"]; a different class is instantiated from the same id (2024-11-18).","keys":[{"file":"config","key":"architectures","before":["NewModel"],"after":["GteModel"]},{"file":"config","key":"model_type","before":"new","after":"gte"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"c9a441e2ab8130e49fb4a623729e555e8575665a","toSha":"40bfdaae8d5499050b9a7b216acb69626463a1f0","commit":"https://huggingface.co/Snowflake/snowflake-arctic-embed-m-v2.0/commit/40bfdaae8d5499050b9a7b216acb69626463a1f0","source":"backfill"},{"id":"Qwen/Qwen2.5-Coder-14B-Instruct-AWQ","date":"2024-11-18T13:04:34.000Z","classes":["tokenizer"],"sentence":"Qwen2.5-Coder-14B-Instruct-AWQ: tokenizer model_max_length 131072 to 32768; truncation defaults change (2024-11-18).","keys":[{"file":"tokenizer","key":"model_max_length","before":131072,"after":32768}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"e37ac9537bb06b98fa5b705fc3c1255881668c03","toSha":"dc6bb0655ea43d158f0760fab0d969a2c1407099","commit":"https://huggingface.co/Qwen/Qwen2.5-Coder-14B-Instruct-AWQ/commit/dc6bb0655ea43d158f0760fab0d969a2c1407099","source":"backfill"},{"id":"Qwen/Qwen2.5-Coder-14B-Instruct","date":"2024-11-18T13:04:23.000Z","classes":["tokenizer"],"sentence":"Qwen2.5-Coder-14B-Instruct: tokenizer model_max_length 131072 to 32768; truncation defaults change (2024-11-18).","keys":[{"file":"tokenizer","key":"model_max_length","before":131072,"after":32768}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"ec98bf0d6c087dca5ef19ee8a10f345c3cd3e0fb","toSha":"7fc9a75b07a3f8f325985b8a10af2e73d7cd63c3","commit":"https://huggingface.co/Qwen/Qwen2.5-Coder-14B-Instruct/commit/7fc9a75b07a3f8f325985b8a10af2e73d7cd63c3","source":"backfill"},{"id":"Qwen/Qwen2.5-Coder-32B-Instruct-AWQ","date":"2024-11-18T12:55:21.000Z","classes":["tokenizer"],"sentence":"Qwen2.5-Coder-32B-Instruct-AWQ: tokenizer model_max_length 131072 to 32768; truncation defaults change (2024-11-18).","keys":[{"file":"tokenizer","key":"model_max_length","before":131072,"after":32768}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"2b4fe968991574e3f76aa95cb4d89b793a7b72de","toSha":"1ed0a6145da0ce550c628e8e8b678f51e695995d","commit":"https://huggingface.co/Qwen/Qwen2.5-Coder-32B-Instruct-AWQ/commit/1ed0a6145da0ce550c628e8e8b678f51e695995d","source":"backfill"},{"id":"Qwen/Qwen2.5-Coder-32B-Instruct","date":"2024-11-18T12:55:03.000Z","classes":["tokenizer"],"sentence":"Qwen2.5-Coder-32B-Instruct: tokenizer model_max_length 131072 to 32768; truncation defaults change (2024-11-18).","keys":[{"file":"tokenizer","key":"model_max_length","before":131072,"after":32768}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"23efe921391d5fbdf76911639850d1d95e8a99d5","toSha":"b47205940b83b5b484577359f71ee7b88472df67","commit":"https://huggingface.co/Qwen/Qwen2.5-Coder-32B-Instruct/commit/b47205940b83b5b484577359f71ee7b88472df67","source":"backfill"},{"id":"Qwen/Qwen2.5-Coder-7B-Instruct-AWQ","date":"2024-11-18T12:54:24.000Z","classes":["tokenizer"],"sentence":"Qwen2.5-Coder-7B-Instruct-AWQ: EOS changed from 151643 to 151645; anything fine-tuned or served against the old value stops at a different token; tokenizer model_max_length 131072 to 32768; truncation defaults change (2024-11-18).","keys":[{"file":"config","key":"eos_token_id","before":151643,"after":151645},{"file":"tokenizer","key":"model_max_length","before":131072,"after":32768}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":{"before":151643,"after":151645},"bos":null},"fromSha":"3f08fd7343ed74290e875bd830961b92b0d62da5","toSha":"8e8ed243bbe6f9a5aff549a0924562fc719b2b8a","commit":"https://huggingface.co/Qwen/Qwen2.5-Coder-7B-Instruct-AWQ/commit/8e8ed243bbe6f9a5aff549a0924562fc719b2b8a","source":"backfill"},{"id":"Qwen/Qwen2.5-Coder-7B-Instruct-GPTQ-Int4","date":"2024-11-18T12:54:13.000Z","classes":["tokenizer"],"sentence":"Qwen2.5-Coder-7B-Instruct-GPTQ-Int4: EOS changed from 151643 to 151645; anything fine-tuned or served against the old value stops at a different token; tokenizer model_max_length 131072 to 32768; truncation defaults change (2024-11-18).","keys":[{"file":"config","key":"eos_token_id","before":151643,"after":151645},{"file":"tokenizer","key":"model_max_length","before":131072,"after":32768}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":{"before":151643,"after":151645},"bos":null},"fromSha":"85fa26bcad08874ed465c540c7b995f247b41f7b","toSha":"3d0e91af138113d0b17d8011adf140d22423d13e","commit":"https://huggingface.co/Qwen/Qwen2.5-Coder-7B-Instruct-GPTQ-Int4/commit/3d0e91af138113d0b17d8011adf140d22423d13e","source":"backfill"},{"id":"Qwen/Qwen2.5-Coder-7B-Instruct","date":"2024-11-18T12:54:07.000Z","classes":["tokenizer"],"sentence":"Qwen2.5-Coder-7B-Instruct: tokenizer model_max_length 131072 to 32768; truncation defaults change (2024-11-18).","keys":[{"file":"tokenizer","key":"model_max_length","before":131072,"after":32768}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"014013f208b0d052dcd0b62bf35efeb573322498","toSha":"0eb6b1ed2d0c4306bc637d09ecef51e59d3dfe05","commit":"https://huggingface.co/Qwen/Qwen2.5-Coder-7B-Instruct/commit/0eb6b1ed2d0c4306bc637d09ecef51e59d3dfe05","source":"backfill"},{"id":"Qwen/Qwen2.5-Coder-7B","date":"2024-11-18T12:54:02.000Z","classes":["tokenizer"],"sentence":"Qwen2.5-Coder-7B: tokenizer model_max_length 131072 to 32768; truncation defaults change (2024-11-18).","keys":[{"file":"tokenizer","key":"model_max_length","before":131072,"after":32768}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"9ec1a913b848a5a70d9fe2261c466a75bdcfeb7c","toSha":"0396a76181e127dfc13e5c5ec48a8cee09938b02","commit":"https://huggingface.co/Qwen/Qwen2.5-Coder-7B/commit/0396a76181e127dfc13e5c5ec48a8cee09938b02","source":"backfill"},{"id":"Qwen/Qwen2.5-Coder-3B-Instruct","date":"2024-11-18T12:53:38.000Z","classes":["tokenizer"],"sentence":"Qwen2.5-Coder-3B-Instruct: tokenizer model_max_length 131072 to 32768; truncation defaults change (2024-11-18).","keys":[{"file":"tokenizer","key":"model_max_length","before":131072,"after":32768}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"2038ec3efe49c590bcb3f3ea984a498eaab814fc","toSha":"89fe5444e8baf5736e70f528f1edcc79e6616ef6","commit":"https://huggingface.co/Qwen/Qwen2.5-Coder-3B-Instruct/commit/89fe5444e8baf5736e70f528f1edcc79e6616ef6","source":"backfill"},{"id":"Qwen/Qwen2.5-Coder-1.5B-Instruct","date":"2024-11-18T12:53:10.000Z","classes":["tokenizer"],"sentence":"Qwen2.5-Coder-1.5B-Instruct: tokenizer model_max_length 131072 to 32768; truncation defaults change (2024-11-18).","keys":[{"file":"tokenizer","key":"model_max_length","before":131072,"after":32768}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"357b899b4714bf46d935fb9911e8139b5b9efc29","toSha":"50285f662d2777f8c585069ed157217de098d26e","commit":"https://huggingface.co/Qwen/Qwen2.5-Coder-1.5B-Instruct/commit/50285f662d2777f8c585069ed157217de098d26e","source":"backfill"},{"id":"Qwen/Qwen2.5-Coder-1.5B","date":"2024-11-18T12:53:05.000Z","classes":["tokenizer"],"sentence":"Qwen2.5-Coder-1.5B: tokenizer model_max_length 131072 to 32768; truncation defaults change (2024-11-18).","keys":[{"file":"tokenizer","key":"model_max_length","before":131072,"after":32768}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"6df6cbb5f18af4eae29e0bc95b7ecd46b07fbdcd","toSha":"df3ce67c0e24480f20468b6ef2894622d69eb73b","commit":"https://huggingface.co/Qwen/Qwen2.5-Coder-1.5B/commit/df3ce67c0e24480f20468b6ef2894622d69eb73b","source":"backfill"},{"id":"Alibaba-NLP/gte-Qwen2-1.5B-instruct","date":"2024-11-15T10:29:12.000Z","classes":["structural","tokenizer"],"sentence":"gte-Qwen2-1.5B-instruct: context window 2048 to 131072 tokens; sliding_window 2048 to 131072; tokenizer model_max_length 2048 to 32768; truncation defaults change (2024-11-15).","keys":[{"file":"config","key":"max_position_embeddings","before":2048,"after":131072},{"file":"config","key":"sliding_window","before":2048,"after":131072},{"file":"tokenizer","key":"model_max_length","before":2048,"after":32768}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":{"before":2048,"after":131072},"eos":null,"bos":null},"fromSha":"3276994ba02b26841920728d1adcf115473c88e9","toSha":"8fdde9cf9d708de7289a58a29877ddf6ae9b618c","commit":"https://huggingface.co/Alibaba-NLP/gte-Qwen2-1.5B-instruct/commit/8fdde9cf9d708de7289a58a29877ddf6ae9b618c","source":"backfill"},{"id":"jinaai/jina-clip-v2","date":"2024-11-12T13:39:37.000Z","classes":["structural"],"sentence":"jina-clip-v2: tie_word_embeddings true removed; the library default decides whether the head is shared; declared dtype \"float16\" to unset (2024-11-12).","keys":[{"file":"config","key":"dtype","before":"float16","after":null},{"file":"config","key":"tie_word_embeddings","before":true,"after":null}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"af015b790b27080f7deb8b9f4b91a5cc70697df4","toSha":"b8b8f72eff90cc350ddaac63084649ef6e459e7a","commit":"https://huggingface.co/jinaai/jina-clip-v2/commit/b8b8f72eff90cc350ddaac63084649ef6e459e7a","source":"backfill"},{"id":"Alibaba-NLP/gte-Qwen2-1.5B-instruct","date":"2024-11-12T08:49:56.000Z","classes":["structural","tokenizer"],"sentence":"gte-Qwen2-1.5B-instruct: context window 131072 to 2048 tokens; sliding_window 131072 to 2048; tokenizer model_max_length 32768 to 2048; truncation defaults change (2024-11-12).","keys":[{"file":"config","key":"max_position_embeddings","before":131072,"after":2048},{"file":"config","key":"sliding_window","before":131072,"after":2048},{"file":"tokenizer","key":"model_max_length","before":32768,"after":2048}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":{"before":131072,"after":2048},"eos":null,"bos":null},"fromSha":"a3fa2698dec82d2860480fa4286382847b58aa12","toSha":"3276994ba02b26841920728d1adcf115473c88e9","commit":"https://huggingface.co/Alibaba-NLP/gte-Qwen2-1.5B-instruct/commit/3276994ba02b26841920728d1adcf115473c88e9","source":"backfill"},{"id":"Qwen/Qwen2.5-Coder-1.5B-Instruct","date":"2024-11-12T02:55:59.000Z","classes":["structural"],"sentence":"Qwen2.5-Coder-1.5B-Instruct: sliding_window 131072 to 32768 (2024-11-12).","keys":[{"file":"config","key":"sliding_window","before":131072,"after":32768}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"253f7f91db978045db0bb3d076a7bd8603708611","toSha":"357b899b4714bf46d935fb9911e8139b5b9efc29","commit":"https://huggingface.co/Qwen/Qwen2.5-Coder-1.5B-Instruct/commit/357b899b4714bf46d935fb9911e8139b5b9efc29","source":"backfill"},{"id":"Qwen/Qwen2.5-Coder-1.5B","date":"2024-11-12T02:55:39.000Z","classes":["structural"],"sentence":"Qwen2.5-Coder-1.5B: sliding_window 131072 to 32768 (2024-11-12).","keys":[{"file":"config","key":"sliding_window","before":131072,"after":32768}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"102a5413ab6b514fb80ebea4931294689f7430d9","toSha":"6df6cbb5f18af4eae29e0bc95b7ecd46b07fbdcd","commit":"https://huggingface.co/Qwen/Qwen2.5-Coder-1.5B/commit/6df6cbb5f18af4eae29e0bc95b7ecd46b07fbdcd","source":"backfill"},{"id":"Qwen/Qwen2.5-Coder-32B-Instruct","date":"2024-11-09T09:33:46.000Z","classes":["tokenizer"],"sentence":"Qwen2.5-Coder-32B-Instruct: EOS changed from 151643 to 151645; anything fine-tuned or served against the old value stops at a different token (2024-11-09).","keys":[{"file":"config","key":"eos_token_id","before":151643,"after":151645}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":{"before":151643,"after":151645},"bos":null},"fromSha":"d3d43e94121feb45a52d52f3613a2f82e17b2977","toSha":"23efe921391d5fbdf76911639850d1d95e8a99d5","commit":"https://huggingface.co/Qwen/Qwen2.5-Coder-32B-Instruct/commit/23efe921391d5fbdf76911639850d1d95e8a99d5","source":"backfill"},{"id":"Qwen/Qwen2.5-Coder-32B-Instruct","date":"2024-11-09T09:33:31.000Z","classes":["generation-defaults"],"sentence":"Qwen2.5-Coder-32B-Instruct: generation defaults changed (repetition_penalty 1.1 to 1.05); a call that does not pass these explicitly now generates differently (2024-11-09).","keys":[{"file":"generation","key":"repetition_penalty","before":1.1,"after":1.05}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"ddff9da45d884bff847142caca6b1be8bad487cc","toSha":"d3d43e94121feb45a52d52f3613a2f82e17b2977","commit":"https://huggingface.co/Qwen/Qwen2.5-Coder-32B-Instruct/commit/d3d43e94121feb45a52d52f3613a2f82e17b2977","source":"backfill"},{"id":"Qwen/Qwen2.5-Coder-14B-Instruct","date":"2024-11-09T09:33:09.000Z","classes":["tokenizer"],"sentence":"Qwen2.5-Coder-14B-Instruct: EOS changed from 151643 to 151645; anything fine-tuned or served against the old value stops at a different token (2024-11-09).","keys":[{"file":"config","key":"eos_token_id","before":151643,"after":151645}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":{"before":151643,"after":151645},"bos":null},"fromSha":"156d6e6c8fc75b368af7a42572ef1caebe25729f","toSha":"f72a6280d9567699c39e4d558eb351289652aa6f","commit":"https://huggingface.co/Qwen/Qwen2.5-Coder-14B-Instruct/commit/f72a6280d9567699c39e4d558eb351289652aa6f","source":"backfill"},{"id":"Qwen/Qwen2.5-Coder-14B-Instruct","date":"2024-11-09T09:32:58.000Z","classes":["generation-defaults"],"sentence":"Qwen2.5-Coder-14B-Instruct: generation defaults changed (repetition_penalty 1.1 to 1.05); a call that does not pass these explicitly now generates differently (2024-11-09).","keys":[{"file":"generation","key":"repetition_penalty","before":1.1,"after":1.05}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"715887aa3a74da2a4d6e9df102a85cb5923a4bca","toSha":"156d6e6c8fc75b368af7a42572ef1caebe25729f","commit":"https://huggingface.co/Qwen/Qwen2.5-Coder-14B-Instruct/commit/156d6e6c8fc75b368af7a42572ef1caebe25729f","source":"backfill"},{"id":"Qwen/Qwen2.5-Coder-7B-Instruct","date":"2024-11-09T09:32:32.000Z","classes":["tokenizer"],"sentence":"Qwen2.5-Coder-7B-Instruct: EOS changed from 151643 to 151645; anything fine-tuned or served against the old value stops at a different token (2024-11-09).","keys":[{"file":"config","key":"eos_token_id","before":151643,"after":151645}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":{"before":151643,"after":151645},"bos":null},"fromSha":"0b044f5762b0be9c48fc2b7bc216454c5eff0bb8","toSha":"014013f208b0d052dcd0b62bf35efeb573322498","commit":"https://huggingface.co/Qwen/Qwen2.5-Coder-7B-Instruct/commit/014013f208b0d052dcd0b62bf35efeb573322498","source":"backfill"},{"id":"Qwen/Qwen2.5-Coder-3B-Instruct","date":"2024-11-09T09:32:12.000Z","classes":["generation-defaults"],"sentence":"Qwen2.5-Coder-3B-Instruct: generation defaults changed (repetition_penalty 1.1 to 1.05); a call that does not pass these explicitly now generates differently (2024-11-09).","keys":[{"file":"generation","key":"repetition_penalty","before":1.1,"after":1.05}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"93d7dcdb8e992d75808533b977e5753266d09b3e","toSha":"2038ec3efe49c590bcb3f3ea984a498eaab814fc","commit":"https://huggingface.co/Qwen/Qwen2.5-Coder-3B-Instruct/commit/2038ec3efe49c590bcb3f3ea984a498eaab814fc","source":"backfill"},{"id":"Qwen/Qwen2.5-Coder-3B-Instruct","date":"2024-11-09T09:31:55.000Z","classes":["tokenizer"],"sentence":"Qwen2.5-Coder-3B-Instruct: EOS changed from 151643 to 151645; anything fine-tuned or served against the old value stops at a different token (2024-11-09).","keys":[{"file":"config","key":"eos_token_id","before":151643,"after":151645}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":{"before":151643,"after":151645},"bos":null},"fromSha":"4083ed00af13fec5298720e9cb005e210f70ac51","toSha":"93d7dcdb8e992d75808533b977e5753266d09b3e","commit":"https://huggingface.co/Qwen/Qwen2.5-Coder-3B-Instruct/commit/93d7dcdb8e992d75808533b977e5753266d09b3e","source":"backfill"},{"id":"Qwen/Qwen2.5-Coder-1.5B-Instruct","date":"2024-11-09T09:31:35.000Z","classes":["tokenizer"],"sentence":"Qwen2.5-Coder-1.5B-Instruct: EOS changed from 151643 to 151645; anything fine-tuned or served against the old value stops at a different token (2024-11-09).","keys":[{"file":"config","key":"eos_token_id","before":151643,"after":151645}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":{"before":151643,"after":151645},"bos":null},"fromSha":"19ddfaeed64b22390fd727ac3caf3c2571fda6e3","toSha":"253f7f91db978045db0bb3d076a7bd8603708611","commit":"https://huggingface.co/Qwen/Qwen2.5-Coder-1.5B-Instruct/commit/253f7f91db978045db0bb3d076a7bd8603708611","source":"backfill"},{"id":"Qwen/Qwen2.5-Coder-3B-Instruct","date":"2024-11-06T11:30:11.000Z","classes":["tokenizer"],"sentence":"Qwen2.5-Coder-3B-Instruct: EOS changed from 151645 to 151643; anything fine-tuned or served against the old value stops at a different token (2024-11-06).","keys":[{"file":"config","key":"eos_token_id","before":151645,"after":151643}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":{"before":151645,"after":151643},"bos":null},"fromSha":"fee34fb68a09e0915a75919cbe86a42a4c543517","toSha":"4083ed00af13fec5298720e9cb005e210f70ac51","commit":"https://huggingface.co/Qwen/Qwen2.5-Coder-3B-Instruct/commit/4083ed00af13fec5298720e9cb005e210f70ac51","source":"backfill"},{"id":"Qwen/Qwen2.5-Coder-3B-Instruct","date":"2024-11-06T11:29:55.000Z","classes":["generation-defaults"],"sentence":"Qwen2.5-Coder-3B-Instruct: generation defaults changed (repetition_penalty 1.05 to 1.1); a call that does not pass these explicitly now generates differently (2024-11-06).","keys":[{"file":"generation","key":"repetition_penalty","before":1.05,"after":1.1}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"098f9cdcafcfa7eceb4827d10427f97c014c811c","toSha":"fee34fb68a09e0915a75919cbe86a42a4c543517","commit":"https://huggingface.co/Qwen/Qwen2.5-Coder-3B-Instruct/commit/fee34fb68a09e0915a75919cbe86a42a4c543517","source":"backfill"},{"id":"Qwen/Qwen2.5-Coder-14B-Instruct","date":"2024-11-06T11:29:37.000Z","classes":["generation-defaults"],"sentence":"Qwen2.5-Coder-14B-Instruct: generation defaults changed (repetition_penalty 1.05 to 1.1); a call that does not pass these explicitly now generates differently (2024-11-06).","keys":[{"file":"generation","key":"repetition_penalty","before":1.05,"after":1.1}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"93e1f0b72787b12b12dbca84a29ff936f9d0ac30","toSha":"715887aa3a74da2a4d6e9df102a85cb5923a4bca","commit":"https://huggingface.co/Qwen/Qwen2.5-Coder-14B-Instruct/commit/715887aa3a74da2a4d6e9df102a85cb5923a4bca","source":"backfill"},{"id":"Qwen/Qwen2.5-Coder-14B-Instruct","date":"2024-11-06T11:29:23.000Z","classes":["tokenizer"],"sentence":"Qwen2.5-Coder-14B-Instruct: EOS changed from 151645 to 151643; anything fine-tuned or served against the old value stops at a different token (2024-11-06).","keys":[{"file":"config","key":"eos_token_id","before":151645,"after":151643}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":{"before":151645,"after":151643},"bos":null},"fromSha":"215507574115bc17fc170a0dc20b86bd7d675d59","toSha":"93e1f0b72787b12b12dbca84a29ff936f9d0ac30","commit":"https://huggingface.co/Qwen/Qwen2.5-Coder-14B-Instruct/commit/93e1f0b72787b12b12dbca84a29ff936f9d0ac30","source":"backfill"},{"id":"Qwen/Qwen2.5-Coder-32B-Instruct","date":"2024-11-06T11:29:11.000Z","classes":["tokenizer"],"sentence":"Qwen2.5-Coder-32B-Instruct: EOS changed from 151645 to 151643; anything fine-tuned or served against the old value stops at a different token (2024-11-06).","keys":[{"file":"config","key":"eos_token_id","before":151645,"after":151643}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":{"before":151645,"after":151643},"bos":null},"fromSha":"a7a975b22db76940f773dca8a5b300ca5865fafa","toSha":"ddff9da45d884bff847142caca6b1be8bad487cc","commit":"https://huggingface.co/Qwen/Qwen2.5-Coder-32B-Instruct/commit/ddff9da45d884bff847142caca6b1be8bad487cc","source":"backfill"},{"id":"Qwen/Qwen2.5-Coder-32B-Instruct","date":"2024-11-06T11:28:48.000Z","classes":["generation-defaults"],"sentence":"Qwen2.5-Coder-32B-Instruct: generation defaults changed (repetition_penalty 1.05 to 1.1); a call that does not pass these explicitly now generates differently (2024-11-06).","keys":[{"file":"generation","key":"repetition_penalty","before":1.05,"after":1.1}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"f447f76906c7759a2b2f9aa05f892361a167c6bb","toSha":"a7a975b22db76940f773dca8a5b300ca5865fafa","commit":"https://huggingface.co/Qwen/Qwen2.5-Coder-32B-Instruct/commit/a7a975b22db76940f773dca8a5b300ca5865fafa","source":"backfill"},{"id":"HuggingFaceTB/SmolLM2-135M-Instruct","date":"2024-10-31T18:03:27.000Z","classes":["template"],"sentence":"SmolLM2-135M-Instruct: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2024-10-31).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"153280e3ff55d19d","after":"872be49dbb638044"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"e682e54cd28bfbd993ed4549dbab7aafbd3e44b7","toSha":"eba0825ddc7730d24a88685a6c0a078cf919b383","commit":"https://huggingface.co/HuggingFaceTB/SmolLM2-135M-Instruct/commit/eba0825ddc7730d24a88685a6c0a078cf919b383","source":"backfill"},{"id":"HuggingFaceTB/SmolLM2-360M-Instruct","date":"2024-10-31T18:02:39.000Z","classes":["template"],"sentence":"SmolLM2-360M-Instruct: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2024-10-31).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"153280e3ff55d19d","after":"872be49dbb638044"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"1717970b6ee66d5bbdc2fa0175463c41eb5b9662","toSha":"51b1c73990e5764f24797c5390c2ca970fb42c1f","commit":"https://huggingface.co/HuggingFaceTB/SmolLM2-360M-Instruct/commit/51b1c73990e5764f24797c5390c2ca970fb42c1f","source":"backfill"},{"id":"HuggingFaceTB/SmolLM2-1.7B-Instruct","date":"2024-10-31T18:02:04.000Z","classes":["template"],"sentence":"SmolLM2-1.7B-Instruct: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2024-10-31).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"153280e3ff55d19d","after":"872be49dbb638044"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"a587c08a0ba91482b3ac0b00c8b50aad514380b8","toSha":"e4f7834c8f5129fc1c8008284905e431cb1bfc40","commit":"https://huggingface.co/HuggingFaceTB/SmolLM2-1.7B-Instruct/commit/e4f7834c8f5129fc1c8008284905e431cb1bfc40","source":"backfill"},{"id":"Qwen/Qwen2.5-7B-Instruct-GPTQ-Int4","date":"2024-10-18T02:59:11.000Z","classes":["template"],"sentence":"Qwen2.5-7B-Instruct-GPTQ-Int4: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2024-10-18).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"e869eb4d8557accc","after":"cd8e9439f0570856"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"d3320b86ad384779c59bd10465547318e1b9d97d","toSha":"e9c932ac1893a49ae0fc497ad6e1e86e2e39af20","commit":"https://huggingface.co/Qwen/Qwen2.5-7B-Instruct-GPTQ-Int4/commit/e9c932ac1893a49ae0fc497ad6e1e86e2e39af20","source":"backfill"},{"id":"Qwen/Qwen2.5-7B-Instruct-GPTQ-Int4","date":"2024-10-18T02:48:27.000Z","classes":["template"],"sentence":"Qwen2.5-7B-Instruct-GPTQ-Int4: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2024-10-18).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"44d5f08f3f72b837","after":"e869eb4d8557accc"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"6e497db5d9b64ab98e8d1278be8ad6added1ac24","toSha":"d3320b86ad384779c59bd10465547318e1b9d97d","commit":"https://huggingface.co/Qwen/Qwen2.5-7B-Instruct-GPTQ-Int4/commit/d3320b86ad384779c59bd10465547318e1b9d97d","source":"backfill"},{"id":"utter-project/EuroLLM-1.7B-Instruct","date":"2024-10-16T10:59:01.000Z","classes":["template"],"sentence":"EuroLLM-1.7B-Instruct: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2024-10-16).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"ef11fb56c1772819","after":"a9092972f456a075"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"fbfaca89e5213c640dbb048ed169ac24213da338","toSha":"cd9ed3768c25b5f71ea70fc3a6f5e7faab349ed8","commit":"https://huggingface.co/utter-project/EuroLLM-1.7B-Instruct/commit/cd9ed3768c25b5f71ea70fc3a6f5e7faab349ed8","source":"backfill"},{"id":"utter-project/EuroLLM-1.7B-Instruct","date":"2024-10-16T09:16:05.000Z","classes":["template"],"sentence":"EuroLLM-1.7B-Instruct: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2024-10-16).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"58c1a1f04baa7ada","after":"ef11fb56c1772819"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"03e76b3df14a077428874fe4f955f1d485c7833a","toSha":"fbfaca89e5213c640dbb048ed169ac24213da338","commit":"https://huggingface.co/utter-project/EuroLLM-1.7B-Instruct/commit/fbfaca89e5213c640dbb048ed169ac24213da338","source":"backfill"},{"id":"utter-project/EuroLLM-1.7B-Instruct","date":"2024-10-16T09:06:53.000Z","classes":["template"],"sentence":"EuroLLM-1.7B-Instruct: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2024-10-16).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"f8da69f2d83ed284","after":"58c1a1f04baa7ada"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"273f29e3ebcaf27c3bd5f6b75e4c62a9eca5f559","toSha":"03e76b3df14a077428874fe4f955f1d485c7833a","commit":"https://huggingface.co/utter-project/EuroLLM-1.7B-Instruct/commit/03e76b3df14a077428874fe4f955f1d485c7833a","source":"backfill"},{"id":"TIGER-Lab/VLM2Vec-Full","date":"2024-10-09T21:15:53.000Z","classes":["structural"],"sentence":"VLM2Vec-Full: hidden size 4096 to 3072; derived parameters 5.63B to 3.82B (-1.81B, -3.62 GB of bf16 weights) (2024-10-09).","keys":[{"file":"config","key":"hidden_size","before":4096,"after":3072}],"consequence":{"params":{"before":5631675712,"after":3821111616},"weightsGbBf16":{"before":11.263,"after":7.642},"trainPeakGb":{"before":396.613,"after":327.308},"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"6f0ed3acf9f55b2acff324cf62833429221a3291","toSha":"d7edfcaee1bc50b334b20aadacbdd9d007ccd904","commit":"https://huggingface.co/TIGER-Lab/VLM2Vec-Full/commit/d7edfcaee1bc50b334b20aadacbdd9d007ccd904","source":"backfill"},{"id":"RedHatAI/Llama-3.2-3B-Instruct-FP8-dynamic","date":"2024-10-09T18:57:15.000Z","classes":["structural"],"sentence":"Llama-3.2-3B-Instruct-FP8-dynamic: quantization_config changed; the stored format of the weights is different (2024-10-09).","keys":[{"file":"config","key":"quantization_config","before":null,"after":{"format":"float-quantized","global_compression_ratio":1.2389887231703325,"kv_cache_scheme":null,"quant_method":"compressed-tensors","quantization_status":"compressed","_rest":"6a23074fcb50e5b0"}}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"8f9fd99bed4ab07341a721a09f787b9bb0c09ca2","toSha":"c308a86de78778c5f904a1d82401ac85e18ca205","commit":"https://huggingface.co/RedHatAI/Llama-3.2-3B-Instruct-FP8-dynamic/commit/c308a86de78778c5f904a1d82401ac85e18ca205","source":"backfill"},{"id":"Qwen/Qwen2.5-1.5B","date":"2024-10-08T10:56:50.000Z","classes":["structural"],"sentence":"Qwen2.5-1.5B: context window 32768 to 131072 tokens; sliding_window 32768 to 131072 (2024-10-08).","keys":[{"file":"config","key":"max_position_embeddings","before":32768,"after":131072},{"file":"config","key":"sliding_window","before":32768,"after":131072}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":{"before":32768,"after":131072},"eos":null,"bos":null},"fromSha":"5f3adf4f8825bd43137c77c3d1db791c197b5bd3","toSha":"8faed761d45a263340a0528343f099c05c9a4323","commit":"https://huggingface.co/Qwen/Qwen2.5-1.5B/commit/8faed761d45a263340a0528343f099c05c9a4323","source":"backfill"},{"id":"openai/whisper-large-v3-turbo","date":"2024-10-01T08:14:50.000Z","classes":["tokenizer"],"sentence":"whisper-large-v3-turbo: pad token unset to \"<|endoftext|>\"; batched training masks a different id (2024-10-01).","keys":[{"file":"tokenizer","key":"pad_token","before":null,"after":"<|endoftext|>"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"57d207fdeb9f44fffde0a0fd30f2cd792df93de5","toSha":"7e2743d4d1fc5182f43d53b0204b2e151ffb6ef7","commit":"https://huggingface.co/openai/whisper-large-v3-turbo/commit/7e2743d4d1fc5182f43d53b0204b2e151ffb6ef7","source":"backfill"},{"id":"openai/whisper-large-v3-turbo","date":"2024-10-01T08:00:34.000Z","classes":["structural"],"sentence":"whisper-large-v3-turbo: declared dtype \"float32\" to \"float16\" (2024-10-01).","keys":[{"file":"config","key":"dtype","before":"float32","after":"float16"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"0631025cc05637b1d1af3efd7b87832707b5b356","toSha":"57d207fdeb9f44fffde0a0fd30f2cd792df93de5","commit":"https://huggingface.co/openai/whisper-large-v3-turbo/commit/57d207fdeb9f44fffde0a0fd30f2cd792df93de5","source":"backfill"},{"id":"utter-project/EuroLLM-1.7B-Instruct","date":"2024-10-01T07:23:02.000Z","classes":["structural"],"sentence":"EuroLLM-1.7B-Instruct: context window 8192 to 4096 tokens (2024-10-01).","keys":[{"file":"config","key":"max_position_embeddings","before":8192,"after":4096}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":{"before":8192,"after":4096},"eos":null,"bos":null},"fromSha":"c3d05511128e1978ddb3ab69f482f0304785d066","toSha":"273f29e3ebcaf27c3bd5f6b75e4c62a9eca5f559","commit":"https://huggingface.co/utter-project/EuroLLM-1.7B-Instruct/commit/273f29e3ebcaf27c3bd5f6b75e4c62a9eca5f559","source":"backfill"},{"id":"unsloth/Llama-3.2-3B-Instruct","date":"2024-09-29T08:23:04.000Z","classes":["tokenizer"],"sentence":"Llama-3.2-3B-Instruct: pad token unset to \"<|finetune_right_pad_id|>\"; batched training masks a different id (2024-09-29).","keys":[{"file":"tokenizer","key":"pad_token","before":null,"after":"<|finetune_right_pad_id|>"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"1c0c6f4320a774461651ff0e95745ac88c97a1e8","toSha":"cfe38f8f38afea8b9f9cc66f634bf53c321cf426","commit":"https://huggingface.co/unsloth/Llama-3.2-3B-Instruct/commit/cfe38f8f38afea8b9f9cc66f634bf53c321cf426","source":"backfill"},{"id":"unsloth/Llama-3.2-1B-Instruct","date":"2024-09-29T08:21:55.000Z","classes":["tokenizer"],"sentence":"Llama-3.2-1B-Instruct: pad token unset to \"<|finetune_right_pad_id|>\"; batched training masks a different id (2024-09-29).","keys":[{"file":"tokenizer","key":"pad_token","before":null,"after":"<|finetune_right_pad_id|>"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"d5a34d2d81ec5c33c511d632517d2cc8b4395792","toSha":"bd6d97814a86488154b2032080b776ca831e5547","commit":"https://huggingface.co/unsloth/Llama-3.2-1B-Instruct/commit/bd6d97814a86488154b2032080b776ca831e5547","source":"backfill"},{"id":"unsloth/Llama-3.2-1B-Instruct","date":"2024-09-29T08:21:17.000Z","classes":["tokenizer"],"sentence":"Llama-3.2-1B-Instruct: pad token \"<|finetune_right_pad_id|>\" to unset; batched training masks a different id (2024-09-29).","keys":[{"file":"tokenizer","key":"pad_token","before":"<|finetune_right_pad_id|>","after":null}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"4e20426873bde3b9bae9ed89fc2f95ed9638d23b","toSha":"d5a34d2d81ec5c33c511d632517d2cc8b4395792","commit":"https://huggingface.co/unsloth/Llama-3.2-1B-Instruct/commit/d5a34d2d81ec5c33c511d632517d2cc8b4395792","source":"backfill"},{"id":"unsloth/Llama-3.2-3B-Instruct","date":"2024-09-29T08:20:40.000Z","classes":["tokenizer"],"sentence":"Llama-3.2-3B-Instruct: pad token \"<|finetune_right_pad_id|>\" to unset; batched training masks a different id (2024-09-29).","keys":[{"file":"tokenizer","key":"pad_token","before":"<|finetune_right_pad_id|>","after":null}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"d374946d800359110f88446ef37fb111cc4269d4","toSha":"1c0c6f4320a774461651ff0e95745ac88c97a1e8","commit":"https://huggingface.co/unsloth/Llama-3.2-3B-Instruct/commit/1c0c6f4320a774461651ff0e95745ac88c97a1e8","source":"backfill"},{"id":"meta-llama/Llama-3.2-1B-Instruct","date":"2024-09-25T15:42:33.000Z","classes":["template"],"sentence":"Llama-3.2-1B-Instruct: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2024-09-25).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"b48c47f644389271","after":"5816fce10444e03c"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"c4219cc9e642e492fd0219283fa3c674804bb8ed","toSha":"0907763fe3a4b2e195d576cbc0144e38cdabdce5","commit":"https://huggingface.co/meta-llama/Llama-3.2-1B-Instruct/commit/0907763fe3a4b2e195d576cbc0144e38cdabdce5","source":"backfill"},{"id":"meta-llama/Llama-3.2-3B-Instruct","date":"2024-09-25T15:42:29.000Z","classes":["template"],"sentence":"Llama-3.2-3B-Instruct: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2024-09-25).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"b48c47f644389271","after":"5816fce10444e03c"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"f6dcb8adea08358576f9a68d351035ee0b730414","toSha":"4e74cfca7240436b64177993bc270c1711093519","commit":"https://huggingface.co/meta-llama/Llama-3.2-3B-Instruct/commit/4e74cfca7240436b64177993bc270c1711093519","source":"backfill"},{"id":"meta-llama/Llama-3.2-3B","date":"2024-09-25T08:45:38.000Z","classes":["generation-defaults"],"sentence":"Llama-3.2-3B: generation_config.json was removed; whatever defaults it set now come from the library (2024-09-25).","keys":[{"file":"generation","key":"(file)","before":"present","after":null}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"1b21fcd6b6dc69aec7af0c8d5099de7f1893df37","toSha":"8d482456ff9cb586a45aa463a3f87537a63ee20f","commit":"https://huggingface.co/meta-llama/Llama-3.2-3B/commit/8d482456ff9cb586a45aa463a3f87537a63ee20f","source":"backfill"},{"id":"meta-llama/Llama-3.2-1B","date":"2024-09-25T08:45:35.000Z","classes":["generation-defaults"],"sentence":"Llama-3.2-1B: generation_config.json was removed; whatever defaults it set now come from the library (2024-09-25).","keys":[{"file":"generation","key":"(file)","before":"present","after":null}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"8a8776e81c703c42c1ed14a8e081543d6336c3e3","toSha":"8a16658ef96719bdf493d7f1816dae790fb57fae","commit":"https://huggingface.co/meta-llama/Llama-3.2-1B/commit/8a16658ef96719bdf493d7f1816dae790fb57fae","source":"backfill"},{"id":"meta-llama/Llama-3.2-1B","date":"2024-09-24T20:22:05.000Z","classes":["structural"],"sentence":"Llama-3.2-1B: rope_scaling values changed ({\"factor\":8,\"high_freq_factor\":4,\"low_freq_factor\":1,\"ori... to {\"factor\":32,\"high_freq_factor\":4,\"low_freq_factor\":1,\"or...); positions are computed differently (2024-09-24).","keys":[{"file":"config","key":"rope_scaling","before":{"factor":8,"high_freq_factor":4,"low_freq_factor":1,"original_max_position_embeddings":8192,"rope_type":"llama3"},"after":{"factor":32,"high_freq_factor":4,"low_freq_factor":1,"original_max_position_embeddings":8192,"rope_type":"llama3"}}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"962c3106fec9c7b4d6cc6faa5de51a1b9aee137c","toSha":"8a8776e81c703c42c1ed14a8e081543d6336c3e3","commit":"https://huggingface.co/meta-llama/Llama-3.2-1B/commit/8a8776e81c703c42c1ed14a8e081543d6336c3e3","source":"backfill"},{"id":"meta-llama/Llama-3.2-3B","date":"2024-09-24T20:21:49.000Z","classes":["structural"],"sentence":"Llama-3.2-3B: rope_scaling values changed ({\"factor\":8,\"high_freq_factor\":4,\"low_freq_factor\":1,\"ori... to {\"factor\":32,\"high_freq_factor\":4,\"low_freq_factor\":1,\"or...); positions are computed differently (2024-09-24).","keys":[{"file":"config","key":"rope_scaling","before":{"factor":8,"high_freq_factor":4,"low_freq_factor":1,"original_max_position_embeddings":8192,"rope_type":"llama3"},"after":{"factor":32,"high_freq_factor":4,"low_freq_factor":1,"original_max_position_embeddings":8192,"rope_type":"llama3"}}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"251c4ef1ac2fe50aa6a0948d742287aaeb9bb159","toSha":"1b21fcd6b6dc69aec7af0c8d5099de7f1893df37","commit":"https://huggingface.co/meta-llama/Llama-3.2-3B/commit/1b21fcd6b6dc69aec7af0c8d5099de7f1893df37","source":"backfill"},{"id":"utter-project/EuroLLM-1.7B-Instruct","date":"2024-09-24T13:25:28.000Z","classes":["structural","template"],"sentence":"EuroLLM-1.7B-Instruct: context window 4096 to 8192 tokens; the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2024-09-24).","keys":[{"file":"config","key":"max_position_embeddings","before":4096,"after":8192},{"file":"tokenizer","key":"chat_template_sha256","before":"58c1a1f04baa7ada","after":"f8da69f2d83ed284"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":{"before":4096,"after":8192},"eos":null,"bos":null},"fromSha":"c00431158750cc7f03f668bd4faec98bb68bc18f","toSha":"c3d05511128e1978ddb3ab69f482f0304785d066","commit":"https://huggingface.co/utter-project/EuroLLM-1.7B-Instruct/commit/c3d05511128e1978ddb3ab69f482f0304785d066","source":"backfill"},{"id":"meta-llama/Llama-3.2-3B-Instruct","date":"2024-09-24T00:23:07.000Z","classes":["structural"],"sentence":"Llama-3.2-3B-Instruct: rope_scaling values changed ({\"factor\":8,\"high_freq_factor\":4,\"low_freq_factor\":1,\"ori... to {\"factor\":32,\"high_freq_factor\":4,\"low_freq_factor\":1,\"or...); positions are computed differently (2024-09-24).","keys":[{"file":"config","key":"rope_scaling","before":{"factor":8,"high_freq_factor":4,"low_freq_factor":1,"original_max_position_embeddings":8192,"rope_type":"llama3"},"after":{"factor":32,"high_freq_factor":4,"low_freq_factor":1,"original_max_position_embeddings":8192,"rope_type":"llama3"}}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"e6c23ae5e6e92a9076abdd4818ac0404cd03af9b","toSha":"f6dcb8adea08358576f9a68d351035ee0b730414","commit":"https://huggingface.co/meta-llama/Llama-3.2-3B-Instruct/commit/f6dcb8adea08358576f9a68d351035ee0b730414","source":"backfill"},{"id":"meta-llama/Llama-3.2-1B-Instruct","date":"2024-09-24T00:22:38.000Z","classes":["structural"],"sentence":"Llama-3.2-1B-Instruct: rope_scaling values changed ({\"factor\":8,\"high_freq_factor\":4,\"low_freq_factor\":1,\"ori... to {\"factor\":32,\"high_freq_factor\":4,\"low_freq_factor\":1,\"or...); positions are computed differently (2024-09-24).","keys":[{"file":"config","key":"rope_scaling","before":{"factor":8,"high_freq_factor":4,"low_freq_factor":1,"original_max_position_embeddings":8192,"rope_type":"llama3"},"after":{"factor":32,"high_freq_factor":4,"low_freq_factor":1,"original_max_position_embeddings":8192,"rope_type":"llama3"}}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"898999bd25b40516fce5a5b8f0948f4c81c650bc","toSha":"c4219cc9e642e492fd0219283fa3c674804bb8ed","commit":"https://huggingface.co/meta-llama/Llama-3.2-1B-Instruct/commit/c4219cc9e642e492fd0219283fa3c674804bb8ed","source":"backfill"},{"id":"boboliu/bge-reranker-v2.5-gemma2-lightweight-gptq","date":"2024-09-23T07:38:47.000Z","classes":["structural"],"sentence":"bge-reranker-v2.5-gemma2-lightweight-gptq: quantization_config changed; the stored format of the weights is different (2024-09-23).","keys":[{"file":"config","key":"quantization_config","before":{"batch_size":1,"bits":4,"block_name_to_quantize":null,"cache_block_outputs":true,"damp_percent":0.1,"dataset":"c4","desc_act":false,"group_size":128,"max_input_length":null,"model_seqlen":null,"module_name_preceding_first_block":null,"modules_in_block_to_quantize":null,"pad_token_id":null,"quant_method":"gptq","sym":true,"tokenizer":null,"true_sequential":true,"use_cuda_fp16":false,"use_exllama":true,"_rest":"345b1d2f73ef8443"},"after":{"batch_size":1,"bits":4,"block_name_to_quantize":null,"cache_block_outputs":true,"damp_percent":0.1,"dataset":"c4","desc_act":false,"group_size":128,"max_input_length":null,"model_seqlen":null,"module_name_preceding_first_block":null,"modules_in_block_to_quantize":null,"pad_token_id":null,"quant_method":"gptq","sym":true,"tokenizer":null,"true_sequential":true,"use_cuda_fp16":false,"use_exllama":false,"_rest":"345b1d2f73ef8443"}}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"3feb704808c8a0221022e97c7a6fe8df9398abc8","toSha":"54e67d19732ed1bed26a008a655d344c087b297f","commit":"https://huggingface.co/boboliu/bge-reranker-v2.5-gemma2-lightweight-gptq/commit/54e67d19732ed1bed26a008a655d344c087b297f","source":"backfill"},{"id":"Qwen/Qwen2.5-Math-1.5B","date":"2024-09-20T07:58:08.000Z","classes":["tokenizer"],"sentence":"Qwen2.5-Math-1.5B: EOS changed from \"<|im_end|>\" to \"<|endoftext|>\"; anything fine-tuned or served against the old value stops at a different token (2024-09-20).","keys":[{"file":"tokenizer","key":"eos_token","before":"<|im_end|>","after":"<|endoftext|>"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":{"before":"<|im_end|>","after":"<|endoftext|>"},"bos":null},"fromSha":"32c01b0b4180405756af1a8e5a1a20ff4436a0ea","toSha":"6d29d576340a5f5e655f56bf95168ec249747903","commit":"https://huggingface.co/Qwen/Qwen2.5-Math-1.5B/commit/6d29d576340a5f5e655f56bf95168ec249747903","source":"backfill"},{"id":"Qwen/Qwen2.5-Coder-7B","date":"2024-09-20T07:58:07.000Z","classes":["tokenizer"],"sentence":"Qwen2.5-Coder-7B: EOS changed from \"<|im_end|>\" to \"<|endoftext|>\"; anything fine-tuned or served against the old value stops at a different token (2024-09-20).","keys":[{"file":"tokenizer","key":"eos_token","before":"<|im_end|>","after":"<|endoftext|>"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":{"before":"<|im_end|>","after":"<|endoftext|>"},"bos":null},"fromSha":"30b6a7e874a78d46b80fa1db3194ea427dd41b08","toSha":"9ec1a913b848a5a70d9fe2261c466a75bdcfeb7c","commit":"https://huggingface.co/Qwen/Qwen2.5-Coder-7B/commit/9ec1a913b848a5a70d9fe2261c466a75bdcfeb7c","source":"backfill"},{"id":"Qwen/Qwen2.5-Coder-1.5B","date":"2024-09-20T07:58:06.000Z","classes":["tokenizer"],"sentence":"Qwen2.5-Coder-1.5B: EOS changed from \"<|im_end|>\" to \"<|endoftext|>\"; anything fine-tuned or served against the old value stops at a different token (2024-09-20).","keys":[{"file":"tokenizer","key":"eos_token","before":"<|im_end|>","after":"<|endoftext|>"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":{"before":"<|im_end|>","after":"<|endoftext|>"},"bos":null},"fromSha":"d3586cfe793730945f8e4d7ef31032a3ee50247d","toSha":"102a5413ab6b514fb80ebea4931294689f7430d9","commit":"https://huggingface.co/Qwen/Qwen2.5-Coder-1.5B/commit/102a5413ab6b514fb80ebea4931294689f7430d9","source":"backfill"},{"id":"Qwen/Qwen2.5-7B","date":"2024-09-20T07:58:01.000Z","classes":["tokenizer"],"sentence":"Qwen2.5-7B: EOS changed from \"<|im_end|>\" to \"<|endoftext|>\"; anything fine-tuned or served against the old value stops at a different token (2024-09-20).","keys":[{"file":"tokenizer","key":"eos_token","before":"<|im_end|>","after":"<|endoftext|>"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":{"before":"<|im_end|>","after":"<|endoftext|>"},"bos":null},"fromSha":"09a0bac5707b43ec44508eab308b0846320c1ed4","toSha":"e25af2efae60472008fbeaf5fb7c4274a87f78d4","commit":"https://huggingface.co/Qwen/Qwen2.5-7B/commit/e25af2efae60472008fbeaf5fb7c4274a87f78d4","source":"backfill"},{"id":"Qwen/Qwen2.5-1.5B","date":"2024-09-20T07:57:58.000Z","classes":["tokenizer"],"sentence":"Qwen2.5-1.5B: EOS changed from \"<|im_end|>\" to \"<|endoftext|>\"; anything fine-tuned or served against the old value stops at a different token (2024-09-20).","keys":[{"file":"tokenizer","key":"eos_token","before":"<|im_end|>","after":"<|endoftext|>"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":{"before":"<|im_end|>","after":"<|endoftext|>"},"bos":null},"fromSha":"e5dfabbcffd9b0c7b31d89b82c5a6b72e663f32c","toSha":"5f3adf4f8825bd43137c77c3d1db791c197b5bd3","commit":"https://huggingface.co/Qwen/Qwen2.5-1.5B/commit/5f3adf4f8825bd43137c77c3d1db791c197b5bd3","source":"backfill"},{"id":"Qwen/Qwen2.5-0.5B","date":"2024-09-20T07:57:57.000Z","classes":["tokenizer"],"sentence":"Qwen2.5-0.5B: EOS changed from \"<|im_end|>\" to \"<|endoftext|>\"; anything fine-tuned or served against the old value stops at a different token (2024-09-20).","keys":[{"file":"tokenizer","key":"eos_token","before":"<|im_end|>","after":"<|endoftext|>"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":{"before":"<|im_end|>","after":"<|endoftext|>"},"bos":null},"fromSha":"2630d3d2321bc1f1878f702166d1b2af019a7310","toSha":"7b382a7278a945cf842898802dd4d862d54c513f","commit":"https://huggingface.co/Qwen/Qwen2.5-0.5B/commit/7b382a7278a945cf842898802dd4d862d54c513f","source":"backfill"},{"id":"Qwen/Qwen2.5-Coder-7B","date":"2024-09-20T03:47:53.000Z","classes":["template"],"sentence":"Qwen2.5-Coder-7B: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2024-09-20).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"44d5f08f3f72b837","after":"cd8e9439f0570856"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"4c1c1611f30619a8695cf6d44b492a25c52b6f00","toSha":"30b6a7e874a78d46b80fa1db3194ea427dd41b08","commit":"https://huggingface.co/Qwen/Qwen2.5-Coder-7B/commit/30b6a7e874a78d46b80fa1db3194ea427dd41b08","source":"backfill"},{"id":"Qwen/Qwen2.5-Coder-1.5B-Instruct","date":"2024-09-20T03:47:19.000Z","classes":["template"],"sentence":"Qwen2.5-Coder-1.5B-Instruct: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2024-09-20).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"44d5f08f3f72b837","after":"cd8e9439f0570856"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"71754d9e15bcc0c2223ad2854682b4ea58fd6f66","toSha":"19ddfaeed64b22390fd727ac3caf3c2571fda6e3","commit":"https://huggingface.co/Qwen/Qwen2.5-Coder-1.5B-Instruct/commit/19ddfaeed64b22390fd727ac3caf3c2571fda6e3","source":"backfill"},{"id":"Qwen/Qwen2.5-Coder-1.5B","date":"2024-09-20T03:47:03.000Z","classes":["template"],"sentence":"Qwen2.5-Coder-1.5B: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2024-09-20).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"44d5f08f3f72b837","after":"cd8e9439f0570856"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"8fb5dcaecbef422c87aa43358f562efa33e6639c","toSha":"d3586cfe793730945f8e4d7ef31032a3ee50247d","commit":"https://huggingface.co/Qwen/Qwen2.5-Coder-1.5B/commit/d3586cfe793730945f8e4d7ef31032a3ee50247d","source":"backfill"},{"id":"Qwen/Qwen2.5-Math-1.5B","date":"2024-09-19T07:30:46.000Z","classes":["template"],"sentence":"Qwen2.5-Math-1.5B: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2024-09-19).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"4302efe46d5ecfe0","after":"c564d12c9977b894"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"37193e87a152563370b983808fa02a8b7d1281bf","toSha":"32c01b0b4180405756af1a8e5a1a20ff4436a0ea","commit":"https://huggingface.co/Qwen/Qwen2.5-Math-1.5B/commit/32c01b0b4180405756af1a8e5a1a20ff4436a0ea","source":"backfill"},{"id":"Qwen/Qwen2.5-Coder-7B-Instruct","date":"2024-09-19T07:30:44.000Z","classes":["template"],"sentence":"Qwen2.5-Coder-7B-Instruct: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2024-09-19).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"d5495a1e5db06111","after":"cd8e9439f0570856"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"d619c0c9d73c20b76470c9bd29f5c435c2f6669d","toSha":"0b044f5762b0be9c48fc2b7bc216454c5eff0bb8","commit":"https://huggingface.co/Qwen/Qwen2.5-Coder-7B-Instruct/commit/0b044f5762b0be9c48fc2b7bc216454c5eff0bb8","source":"backfill"},{"id":"Qwen/Qwen2.5-Coder-7B","date":"2024-09-19T07:30:43.000Z","classes":["template"],"sentence":"Qwen2.5-Coder-7B: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2024-09-19).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"89d726e7795b1019","after":"44d5f08f3f72b837"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"f07775da2cd85e4ab99da077ee4b300f051c02cd","toSha":"4c1c1611f30619a8695cf6d44b492a25c52b6f00","commit":"https://huggingface.co/Qwen/Qwen2.5-Coder-7B/commit/4c1c1611f30619a8695cf6d44b492a25c52b6f00","source":"backfill"},{"id":"Qwen/Qwen2.5-Coder-1.5B-Instruct","date":"2024-09-19T07:30:41.000Z","classes":["template"],"sentence":"Qwen2.5-Coder-1.5B-Instruct: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2024-09-19).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"89d726e7795b1019","after":"44d5f08f3f72b837"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"9ce57ed54f45f61e349bb44ec689e4a013b8be7c","toSha":"71754d9e15bcc0c2223ad2854682b4ea58fd6f66","commit":"https://huggingface.co/Qwen/Qwen2.5-Coder-1.5B-Instruct/commit/71754d9e15bcc0c2223ad2854682b4ea58fd6f66","source":"backfill"},{"id":"Qwen/Qwen2.5-Coder-1.5B","date":"2024-09-19T07:30:40.000Z","classes":["template"],"sentence":"Qwen2.5-Coder-1.5B: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2024-09-19).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"89d726e7795b1019","after":"44d5f08f3f72b837"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"6a7051e60c0bb99f8e4eda1e6c9ad3189e8e09da","toSha":"8fb5dcaecbef422c87aa43358f562efa33e6639c","commit":"https://huggingface.co/Qwen/Qwen2.5-Coder-1.5B/commit/8fb5dcaecbef422c87aa43358f562efa33e6639c","source":"backfill"},{"id":"Qwen/Qwen2.5-72B-Instruct-AWQ","date":"2024-09-19T07:30:38.000Z","classes":["template"],"sentence":"Qwen2.5-72B-Instruct-AWQ: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2024-09-19).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"d5495a1e5db06111","after":"cd8e9439f0570856"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"bfe0f13f8beda9b6c713352fe8377d0fac27116e","toSha":"035adeffb861e8ee98344e2c7d8f50fe585e373d","commit":"https://huggingface.co/Qwen/Qwen2.5-72B-Instruct-AWQ/commit/035adeffb861e8ee98344e2c7d8f50fe585e373d","source":"backfill"},{"id":"Qwen/Qwen2.5-72B-Instruct","date":"2024-09-19T07:30:33.000Z","classes":["template"],"sentence":"Qwen2.5-72B-Instruct: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2024-09-19).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"d5495a1e5db06111","after":"cd8e9439f0570856"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"c970f6d046b6df37baa3d7466df2287fa7c9620e","toSha":"a13fff9ad76700c7ecff2769f75943ba8395b4a7","commit":"https://huggingface.co/Qwen/Qwen2.5-72B-Instruct/commit/a13fff9ad76700c7ecff2769f75943ba8395b4a7","source":"backfill"},{"id":"Qwen/Qwen2.5-32B-Instruct-AWQ","date":"2024-09-19T07:30:29.000Z","classes":["template"],"sentence":"Qwen2.5-32B-Instruct-AWQ: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2024-09-19).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"d5495a1e5db06111","after":"cd8e9439f0570856"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"e5b80c95c8f2b28321d16bb537e60d1fbddd891d","toSha":"4b99fe578fcefa565d7497b70939b0e40de1b434","commit":"https://huggingface.co/Qwen/Qwen2.5-32B-Instruct-AWQ/commit/4b99fe578fcefa565d7497b70939b0e40de1b434","source":"backfill"},{"id":"Qwen/Qwen2.5-32B-Instruct-GPTQ-Int4","date":"2024-09-19T07:30:26.000Z","classes":["template"],"sentence":"Qwen2.5-32B-Instruct-GPTQ-Int4: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2024-09-19).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"d5495a1e5db06111","after":"cd8e9439f0570856"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"8604ddb8831d15cd5d3806ecf307de0f7a0aa09c","toSha":"34f0b153f75fea7e00a683b7505f51f7fd08da25","commit":"https://huggingface.co/Qwen/Qwen2.5-32B-Instruct-GPTQ-Int4/commit/34f0b153f75fea7e00a683b7505f51f7fd08da25","source":"backfill"},{"id":"Qwen/Qwen2.5-32B-Instruct","date":"2024-09-19T07:30:25.000Z","classes":["template"],"sentence":"Qwen2.5-32B-Instruct: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2024-09-19).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"d5495a1e5db06111","after":"cd8e9439f0570856"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"e47090056c773456c2bfdbf5ceb66ed7a2ee9954","toSha":"70e8dfb9ad18a7d499f765fe206ff065ed8ca197","commit":"https://huggingface.co/Qwen/Qwen2.5-32B-Instruct/commit/70e8dfb9ad18a7d499f765fe206ff065ed8ca197","source":"backfill"},{"id":"Qwen/Qwen2.5-14B-Instruct-AWQ","date":"2024-09-19T07:30:22.000Z","classes":["template"],"sentence":"Qwen2.5-14B-Instruct-AWQ: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2024-09-19).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"d5495a1e5db06111","after":"cd8e9439f0570856"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"302d7ea6b811c4c5c6e127a0435f6342dbe9ec72","toSha":"15caf3706c437f72fe4a0b64b4aee94b5e823e9c","commit":"https://huggingface.co/Qwen/Qwen2.5-14B-Instruct-AWQ/commit/15caf3706c437f72fe4a0b64b4aee94b5e823e9c","source":"backfill"},{"id":"Qwen/Qwen2.5-14B-Instruct","date":"2024-09-19T07:30:18.000Z","classes":["template"],"sentence":"Qwen2.5-14B-Instruct: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2024-09-19).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"d5495a1e5db06111","after":"cd8e9439f0570856"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"76ca6ba7b68f0cc64f2e888467ab9c0ba671f4e3","toSha":"502e5d8bfd665ed113fd9b3626445ca7b0596303","commit":"https://huggingface.co/Qwen/Qwen2.5-14B-Instruct/commit/502e5d8bfd665ed113fd9b3626445ca7b0596303","source":"backfill"},{"id":"Qwen/Qwen2.5-7B-Instruct-AWQ","date":"2024-09-19T07:30:15.000Z","classes":["template"],"sentence":"Qwen2.5-7B-Instruct-AWQ: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2024-09-19).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"d5495a1e5db06111","after":"cd8e9439f0570856"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"d20af0af7b0d798e51915c886e8b2cb836da8e7a","toSha":"617b26a2e255b3d75afdfd7f390e7df70b00167c","commit":"https://huggingface.co/Qwen/Qwen2.5-7B-Instruct-AWQ/commit/617b26a2e255b3d75afdfd7f390e7df70b00167c","source":"backfill"},{"id":"Qwen/Qwen2.5-7B-Instruct-GPTQ-Int4","date":"2024-09-19T07:30:12.000Z","classes":["template"],"sentence":"Qwen2.5-7B-Instruct-GPTQ-Int4: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2024-09-19).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"89d726e7795b1019","after":"44d5f08f3f72b837"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"22a88496da3086d81cd665478d03e8a25883aa74","toSha":"6e497db5d9b64ab98e8d1278be8ad6added1ac24","commit":"https://huggingface.co/Qwen/Qwen2.5-7B-Instruct-GPTQ-Int4/commit/6e497db5d9b64ab98e8d1278be8ad6added1ac24","source":"backfill"},{"id":"Qwen/Qwen2.5-7B-Instruct","date":"2024-09-19T07:30:10.000Z","classes":["template"],"sentence":"Qwen2.5-7B-Instruct: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2024-09-19).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"d5495a1e5db06111","after":"cd8e9439f0570856"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"a306ca65c69cb09b9ce8ae8396fd529d509d872e","toSha":"acbd96531cda22292a3ceaa67e984955d3965282","commit":"https://huggingface.co/Qwen/Qwen2.5-7B-Instruct/commit/acbd96531cda22292a3ceaa67e984955d3965282","source":"backfill"},{"id":"Qwen/Qwen2.5-7B","date":"2024-09-19T07:30:09.000Z","classes":["template"],"sentence":"Qwen2.5-7B: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2024-09-19).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"89d726e7795b1019","after":"44d5f08f3f72b837"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"3115c99c189e698dd7274e8a463a1025b0c878a6","toSha":"09a0bac5707b43ec44508eab308b0846320c1ed4","commit":"https://huggingface.co/Qwen/Qwen2.5-7B/commit/09a0bac5707b43ec44508eab308b0846320c1ed4","source":"backfill"},{"id":"Qwen/Qwen2.5-3B-Instruct-AWQ","date":"2024-09-19T07:30:08.000Z","classes":["template"],"sentence":"Qwen2.5-3B-Instruct-AWQ: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2024-09-19).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"d5495a1e5db06111","after":"cd8e9439f0570856"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"a70cbfca87ec9f398f17c3f689d5eaddf2b35dcd","toSha":"a2be73dcf1f9d5172f1f37762ce9ddbfa96e854e","commit":"https://huggingface.co/Qwen/Qwen2.5-3B-Instruct-AWQ/commit/a2be73dcf1f9d5172f1f37762ce9ddbfa96e854e","source":"backfill"},{"id":"Qwen/Qwen2.5-3B-Instruct","date":"2024-09-19T07:30:05.000Z","classes":["template"],"sentence":"Qwen2.5-3B-Instruct: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2024-09-19).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"d5495a1e5db06111","after":"cd8e9439f0570856"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"9fbe4b579dd1096693a521a055df1472cbb211bf","toSha":"82f42baa094a9600e39ccd80d34058aeeb3abbc1","commit":"https://huggingface.co/Qwen/Qwen2.5-3B-Instruct/commit/82f42baa094a9600e39ccd80d34058aeeb3abbc1","source":"backfill"},{"id":"Qwen/Qwen2.5-1.5B-Instruct","date":"2024-09-19T07:29:59.000Z","classes":["template"],"sentence":"Qwen2.5-1.5B-Instruct: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2024-09-19).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"d5495a1e5db06111","after":"cd8e9439f0570856"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"b9d14db30f3849a06f43e2c7d5d413223919f5ed","toSha":"5fee7c4ed634dc66c6e318c8ac2897b8b9154536","commit":"https://huggingface.co/Qwen/Qwen2.5-1.5B-Instruct/commit/5fee7c4ed634dc66c6e318c8ac2897b8b9154536","source":"backfill"},{"id":"Qwen/Qwen2.5-1.5B","date":"2024-09-19T07:29:57.000Z","classes":["template"],"sentence":"Qwen2.5-1.5B: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2024-09-19).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"89d726e7795b1019","after":"44d5f08f3f72b837"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"033b7782da89bcc457997494478bba1bf19025bc","toSha":"e5dfabbcffd9b0c7b31d89b82c5a6b72e663f32c","commit":"https://huggingface.co/Qwen/Qwen2.5-1.5B/commit/e5dfabbcffd9b0c7b31d89b82c5a6b72e663f32c","source":"backfill"},{"id":"Qwen/Qwen2.5-0.5B-Instruct","date":"2024-09-19T07:29:52.000Z","classes":["template"],"sentence":"Qwen2.5-0.5B-Instruct: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2024-09-19).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"d5495a1e5db06111","after":"cd8e9439f0570856"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"a338b55dd21219a5f4da42bc11a9313d1a27d4cc","toSha":"a8b602d9dafd3a75d382e62757d83d89fca3be54","commit":"https://huggingface.co/Qwen/Qwen2.5-0.5B-Instruct/commit/a8b602d9dafd3a75d382e62757d83d89fca3be54","source":"backfill"},{"id":"Qwen/Qwen2.5-0.5B","date":"2024-09-19T07:29:50.000Z","classes":["template"],"sentence":"Qwen2.5-0.5B: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2024-09-19).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"89d726e7795b1019","after":"44d5f08f3f72b837"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"74e0d9cfa1ba6f779d35f63aeb6cea7511937269","toSha":"2630d3d2321bc1f1878f702166d1b2af019a7310","commit":"https://huggingface.co/Qwen/Qwen2.5-0.5B/commit/2630d3d2321bc1f1878f702166d1b2af019a7310","source":"backfill"},{"id":"Qwen/Qwen2.5-Coder-1.5B-Instruct","date":"2024-09-19T04:01:58.000Z","classes":["tokenizer","generation-defaults"],"sentence":"Qwen2.5-Coder-1.5B-Instruct: EOS changed from 151643 to [151645,151643]; anything fine-tuned or served against the old value stops at a different token; pad token unset to 151643; batched training masks a different id; generation defaults changed (do_sample unset to true, max_new_tokens 2048 to unset, repetition_penalty unset to 1.1, temperature unset to 0.7); a call that does not pass these explicitly now generates differently (2024-09-19).","keys":[{"file":"generation","key":"do_sample","before":null,"after":true},{"file":"generation","key":"eos_token_id","before":151643,"after":[151645,151643]},{"file":"generation","key":"max_new_tokens","before":2048,"after":null},{"file":"generation","key":"pad_token_id","before":null,"after":151643},{"file":"generation","key":"repetition_penalty","before":null,"after":1.1},{"file":"generation","key":"temperature","before":null,"after":0.7},{"file":"generation","key":"top_k","before":null,"after":20},{"file":"generation","key":"top_p","before":null,"after":0.8}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":{"before":151643,"after":[151645,151643]},"bos":null},"fromSha":"098145f275b49d4517571a8c5d1e7896f68797d8","toSha":"9ce57ed54f45f61e349bb44ec689e4a013b8be7c","commit":"https://huggingface.co/Qwen/Qwen2.5-Coder-1.5B-Instruct/commit/9ce57ed54f45f61e349bb44ec689e4a013b8be7c","source":"backfill"},{"id":"Qwen/Qwen2.5-Coder-7B-Instruct","date":"2024-09-18T14:59:22.000Z","classes":["structural"],"sentence":"Qwen2.5-Coder-7B-Instruct: context window 131072 to 32768 tokens; rope_scaling removed (was {\"factor\":4,\"original_max_position_embeddings\":32768,\"typ...); positions are no longer scaled (2024-09-18).","keys":[{"file":"config","key":"max_position_embeddings","before":131072,"after":32768},{"file":"config","key":"rope_scaling","before":{"factor":4,"original_max_position_embeddings":32768,"type":"yarn"},"after":null}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":{"before":131072,"after":32768},"eos":null,"bos":null},"fromSha":"8d20e2b046afe44cb4f759cff10574b995c0ddb8","toSha":"d619c0c9d73c20b76470c9bd29f5c435c2f6669d","commit":"https://huggingface.co/Qwen/Qwen2.5-Coder-7B-Instruct/commit/d619c0c9d73c20b76470c9bd29f5c435c2f6669d","source":"backfill"},{"id":"Qwen/Qwen2.5-Coder-7B","date":"2024-09-18T14:58:37.000Z","classes":["structural"],"sentence":"Qwen2.5-Coder-7B: context window 131072 to 32768 tokens; rope_scaling removed (was {\"factor\":4,\"original_max_position_embeddings\":32768,\"typ...); positions are no longer scaled (2024-09-18).","keys":[{"file":"config","key":"max_position_embeddings","before":131072,"after":32768},{"file":"config","key":"rope_scaling","before":{"factor":4,"original_max_position_embeddings":32768,"type":"yarn"},"after":null}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":{"before":131072,"after":32768},"eos":null,"bos":null},"fromSha":"f52e93b2dd070f1c38fd0c403634261052d5c0fb","toSha":"f07775da2cd85e4ab99da077ee4b300f051c02cd","commit":"https://huggingface.co/Qwen/Qwen2.5-Coder-7B/commit/f07775da2cd85e4ab99da077ee4b300f051c02cd","source":"backfill"},{"id":"Qwen/Qwen2.5-Coder-1.5B-Instruct","date":"2024-09-18T14:57:32.000Z","classes":["structural"],"sentence":"Qwen2.5-Coder-1.5B-Instruct: sliding_window 32768 to 131072 (2024-09-18).","keys":[{"file":"config","key":"sliding_window","before":32768,"after":131072}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"e3b23484d3a9a85090a66068cd8396cc015c3637","toSha":"098145f275b49d4517571a8c5d1e7896f68797d8","commit":"https://huggingface.co/Qwen/Qwen2.5-Coder-1.5B-Instruct/commit/098145f275b49d4517571a8c5d1e7896f68797d8","source":"backfill"},{"id":"Qwen/Qwen2.5-Coder-1.5B","date":"2024-09-18T14:56:42.000Z","classes":["structural"],"sentence":"Qwen2.5-Coder-1.5B: sliding_window 32768 to 131072 (2024-09-18).","keys":[{"file":"config","key":"sliding_window","before":32768,"after":131072}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"94fddf431e34febf6c1e297d26d30c95c4f43262","toSha":"6a7051e60c0bb99f8e4eda1e6c9ad3189e8e09da","commit":"https://huggingface.co/Qwen/Qwen2.5-Coder-1.5B/commit/6a7051e60c0bb99f8e4eda1e6c9ad3189e8e09da","source":"backfill"},{"id":"Qwen/Qwen2.5-7B-Instruct-GPTQ-Int4","date":"2024-09-18T08:29:42.000Z","classes":["structural"],"sentence":"Qwen2.5-7B-Instruct-GPTQ-Int4: context window 8192 to 32768 tokens; rope_scaling removed (was {\"factor\":4,\"original_max_position_embeddings\":32768,\"typ...); positions are no longer scaled; sliding_window 32768 to 131072 (2024-09-18).","keys":[{"file":"config","key":"max_position_embeddings","before":8192,"after":32768},{"file":"config","key":"rope_scaling","before":{"factor":4,"original_max_position_embeddings":32768,"type":"yarn"},"after":null},{"file":"config","key":"sliding_window","before":32768,"after":131072}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":{"before":8192,"after":32768},"eos":null,"bos":null},"fromSha":"4014efd717cc0459fff730fe76b7c32b114e3d14","toSha":"22a88496da3086d81cd665478d03e8a25883aa74","commit":"https://huggingface.co/Qwen/Qwen2.5-7B-Instruct-GPTQ-Int4/commit/22a88496da3086d81cd665478d03e8a25883aa74","source":"backfill"},{"id":"Qwen/Qwen2.5-72B-Instruct-AWQ","date":"2024-09-18T06:30:06.000Z","classes":["structural"],"sentence":"Qwen2.5-72B-Instruct-AWQ: rope_scaling removed (was {\"factor\":4,\"original_max_position_embeddings\":32768,\"typ...); positions are no longer scaled; sliding_window unset to 131072 (2024-09-18).","keys":[{"file":"config","key":"rope_scaling","before":{"factor":4,"original_max_position_embeddings":32768,"type":"yarn"},"after":null},{"file":"config","key":"sliding_window","before":null,"after":131072}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"026801bff914bffb930a465ab1ac4512bd9ca8b7","toSha":"bfe0f13f8beda9b6c713352fe8377d0fac27116e","commit":"https://huggingface.co/Qwen/Qwen2.5-72B-Instruct-AWQ/commit/bfe0f13f8beda9b6c713352fe8377d0fac27116e","source":"backfill"},{"id":"Qwen/Qwen2.5-32B-Instruct-AWQ","date":"2024-09-18T06:28:26.000Z","classes":["structural"],"sentence":"Qwen2.5-32B-Instruct-AWQ: sliding_window 32768 to 131072 (2024-09-18).","keys":[{"file":"config","key":"sliding_window","before":32768,"after":131072}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"14d7f9f6fe0d89f56b552f2cd34f3c6c9a4cbb04","toSha":"e5b80c95c8f2b28321d16bb537e60d1fbddd891d","commit":"https://huggingface.co/Qwen/Qwen2.5-32B-Instruct-AWQ/commit/e5b80c95c8f2b28321d16bb537e60d1fbddd891d","source":"backfill"},{"id":"Qwen/Qwen2.5-14B-Instruct-AWQ","date":"2024-09-18T06:27:37.000Z","classes":["structural"],"sentence":"Qwen2.5-14B-Instruct-AWQ: sliding_window 32768 to 131072 (2024-09-18).","keys":[{"file":"config","key":"sliding_window","before":32768,"after":131072}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"25d0decbe26162fe135c517dc28c1fa1e66f981b","toSha":"302d7ea6b811c4c5c6e127a0435f6342dbe9ec72","commit":"https://huggingface.co/Qwen/Qwen2.5-14B-Instruct-AWQ/commit/302d7ea6b811c4c5c6e127a0435f6342dbe9ec72","source":"backfill"},{"id":"Qwen/Qwen2.5-7B-Instruct-AWQ","date":"2024-09-18T05:48:38.000Z","classes":["structural"],"sentence":"Qwen2.5-7B-Instruct-AWQ: rope_scaling removed (was {\"factor\":4,\"original_max_position_embeddings\":32768,\"typ...); positions are no longer scaled; sliding_window 32768 to 131072 (2024-09-18).","keys":[{"file":"config","key":"rope_scaling","before":{"factor":4,"original_max_position_embeddings":32768,"type":"yarn"},"after":null},{"file":"config","key":"sliding_window","before":32768,"after":131072}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"73f22df4e9ab5b6b757255cf08ef05f51f20fa90","toSha":"d20af0af7b0d798e51915c886e8b2cb836da8e7a","commit":"https://huggingface.co/Qwen/Qwen2.5-7B-Instruct-AWQ/commit/d20af0af7b0d798e51915c886e8b2cb836da8e7a","source":"backfill"},{"id":"Qwen/Qwen2.5-7B-Instruct","date":"2024-09-18T05:48:17.000Z","classes":["structural"],"sentence":"Qwen2.5-7B-Instruct: context window 131072 to 32768 tokens; rope_scaling removed (was {\"factor\":4,\"original_max_position_embeddings\":32768,\"typ...); positions are no longer scaled (2024-09-18).","keys":[{"file":"config","key":"max_position_embeddings","before":131072,"after":32768},{"file":"config","key":"rope_scaling","before":{"factor":4,"original_max_position_embeddings":32768,"type":"yarn"},"after":null}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":{"before":131072,"after":32768},"eos":null,"bos":null},"fromSha":"929d90a7473be218ae132181708d94dbf08a0420","toSha":"a306ca65c69cb09b9ce8ae8396fd529d509d872e","commit":"https://huggingface.co/Qwen/Qwen2.5-7B-Instruct/commit/a306ca65c69cb09b9ce8ae8396fd529d509d872e","source":"backfill"},{"id":"Qwen/Qwen2.5-72B-Instruct","date":"2024-09-18T05:47:19.000Z","classes":["structural"],"sentence":"Qwen2.5-72B-Instruct: context window 131072 to 32768 tokens; rope_scaling removed (was {\"factor\":4,\"original_max_position_embeddings\":32768,\"typ...); positions are no longer scaled (2024-09-18).","keys":[{"file":"config","key":"max_position_embeddings","before":131072,"after":32768},{"file":"config","key":"rope_scaling","before":{"factor":4,"original_max_position_embeddings":32768,"type":"yarn"},"after":null}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":{"before":131072,"after":32768},"eos":null,"bos":null},"fromSha":"f073433cb484002b27d7a84e8bce1c7435e14a1c","toSha":"c970f6d046b6df37baa3d7466df2287fa7c9620e","commit":"https://huggingface.co/Qwen/Qwen2.5-72B-Instruct/commit/c970f6d046b6df37baa3d7466df2287fa7c9620e","source":"backfill"},{"id":"Qwen/Qwen2.5-32B-Instruct-GPTQ-Int4","date":"2024-09-18T05:45:35.000Z","classes":["structural"],"sentence":"Qwen2.5-32B-Instruct-GPTQ-Int4: rope_scaling removed (was {\"factor\":4,\"original_max_position_embeddings\":32768,\"typ...); positions are no longer scaled; sliding_window 32768 to 131072 (2024-09-18).","keys":[{"file":"config","key":"rope_scaling","before":{"factor":4,"original_max_position_embeddings":32768,"type":"yarn"},"after":null},{"file":"config","key":"sliding_window","before":32768,"after":131072}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"9239191be9d2202d2018e2ed73e37b644b4b3816","toSha":"8604ddb8831d15cd5d3806ecf307de0f7a0aa09c","commit":"https://huggingface.co/Qwen/Qwen2.5-32B-Instruct-GPTQ-Int4/commit/8604ddb8831d15cd5d3806ecf307de0f7a0aa09c","source":"backfill"},{"id":"Qwen/Qwen2.5-32B-Instruct","date":"2024-09-18T05:45:11.000Z","classes":["structural"],"sentence":"Qwen2.5-32B-Instruct: context window 131072 to 32768 tokens; rope_scaling removed (was {\"factor\":4,\"original_max_position_embeddings\":32768,\"typ...); positions are no longer scaled (2024-09-18).","keys":[{"file":"config","key":"max_position_embeddings","before":131072,"after":32768},{"file":"config","key":"rope_scaling","before":{"factor":4,"original_max_position_embeddings":32768,"type":"yarn"},"after":null}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":{"before":131072,"after":32768},"eos":null,"bos":null},"fromSha":"a7e76d7ea0d64956a3238b461efa58a47f77c8b3","toSha":"e47090056c773456c2bfdbf5ceb66ed7a2ee9954","commit":"https://huggingface.co/Qwen/Qwen2.5-32B-Instruct/commit/e47090056c773456c2bfdbf5ceb66ed7a2ee9954","source":"backfill"},{"id":"Qwen/Qwen2.5-14B-Instruct","date":"2024-09-18T05:44:12.000Z","classes":["structural"],"sentence":"Qwen2.5-14B-Instruct: context window 131072 to 32768 tokens; rope_scaling removed (was {\"factor\":4,\"original_max_position_embeddings\":32768,\"typ...); positions are no longer scaled; sliding_window 32768 to 131072 (2024-09-18).","keys":[{"file":"config","key":"max_position_embeddings","before":131072,"after":32768},{"file":"config","key":"rope_scaling","before":{"factor":4,"original_max_position_embeddings":32768,"type":"yarn"},"after":null},{"file":"config","key":"sliding_window","before":32768,"after":131072}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":{"before":131072,"after":32768},"eos":null,"bos":null},"fromSha":"e648520a381cc1bac5b91ff5f42879ba32550fc9","toSha":"76ca6ba7b68f0cc64f2e888467ab9c0ba671f4e3","commit":"https://huggingface.co/Qwen/Qwen2.5-14B-Instruct/commit/76ca6ba7b68f0cc64f2e888467ab9c0ba671f4e3","source":"backfill"},{"id":"Qwen/Qwen2.5-Coder-7B-Instruct","date":"2024-09-18T03:28:39.000Z","classes":["template"],"sentence":"Qwen2.5-Coder-7B-Instruct: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2024-09-18).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"89d726e7795b1019","after":"d5495a1e5db06111"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"6ab006b7202ec68b6df47693b5c3003998dc3124","toSha":"8d20e2b046afe44cb4f759cff10574b995c0ddb8","commit":"https://huggingface.co/Qwen/Qwen2.5-Coder-7B-Instruct/commit/8d20e2b046afe44cb4f759cff10574b995c0ddb8","source":"backfill"},{"id":"Qwen/Qwen2.5-32B-Instruct-GPTQ-Int4","date":"2024-09-18T02:13:21.000Z","classes":["template"],"sentence":"Qwen2.5-32B-Instruct-GPTQ-Int4: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2024-09-18).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"89d726e7795b1019","after":"d5495a1e5db06111"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"3fa9c79b1c7afa3aff71daed487ef6af89be942c","toSha":"9239191be9d2202d2018e2ed73e37b644b4b3816","commit":"https://huggingface.co/Qwen/Qwen2.5-32B-Instruct-GPTQ-Int4/commit/9239191be9d2202d2018e2ed73e37b644b4b3816","source":"backfill"},{"id":"Qwen/Qwen2.5-32B-Instruct","date":"2024-09-18T01:45:12.000Z","classes":["template"],"sentence":"Qwen2.5-32B-Instruct: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2024-09-18).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"89d726e7795b1019","after":"d5495a1e5db06111"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"59eaef4212cfe2f5df3cb438c9acd0934913d8a7","toSha":"a7e76d7ea0d64956a3238b461efa58a47f77c8b3","commit":"https://huggingface.co/Qwen/Qwen2.5-32B-Instruct/commit/a7e76d7ea0d64956a3238b461efa58a47f77c8b3","source":"backfill"},{"id":"Qwen/Qwen2.5-14B-Instruct","date":"2024-09-18T01:43:14.000Z","classes":["template"],"sentence":"Qwen2.5-14B-Instruct: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2024-09-18).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"89d726e7795b1019","after":"d5495a1e5db06111"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"336cd7ff9b3a3f4392a6172197cdcfd7a9d73a30","toSha":"e648520a381cc1bac5b91ff5f42879ba32550fc9","commit":"https://huggingface.co/Qwen/Qwen2.5-14B-Instruct/commit/e648520a381cc1bac5b91ff5f42879ba32550fc9","source":"backfill"},{"id":"Qwen/Qwen2.5-7B-Instruct","date":"2024-09-18T01:42:04.000Z","classes":["template"],"sentence":"Qwen2.5-7B-Instruct: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2024-09-18).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"89d726e7795b1019","after":"d5495a1e5db06111"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"605f1116cfdfd9f5d5eb7d3f9ee2c546d1b1a0e3","toSha":"929d90a7473be218ae132181708d94dbf08a0420","commit":"https://huggingface.co/Qwen/Qwen2.5-7B-Instruct/commit/929d90a7473be218ae132181708d94dbf08a0420","source":"backfill"},{"id":"Qwen/Qwen2.5-3B-Instruct","date":"2024-09-18T01:40:42.000Z","classes":["template"],"sentence":"Qwen2.5-3B-Instruct: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2024-09-18).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"89d726e7795b1019","after":"d5495a1e5db06111"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"3ca5f5bff0f44b2b436a9aba99b900ab157c9a27","toSha":"9fbe4b579dd1096693a521a055df1472cbb211bf","commit":"https://huggingface.co/Qwen/Qwen2.5-3B-Instruct/commit/9fbe4b579dd1096693a521a055df1472cbb211bf","source":"backfill"},{"id":"Qwen/Qwen2.5-1.5B-Instruct","date":"2024-09-18T01:39:57.000Z","classes":["template"],"sentence":"Qwen2.5-1.5B-Instruct: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2024-09-18).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"89d726e7795b1019","after":"d5495a1e5db06111"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"775b11afaf83e0dc75bd5abaf90133e47b3ec082","toSha":"b9d14db30f3849a06f43e2c7d5d413223919f5ed","commit":"https://huggingface.co/Qwen/Qwen2.5-1.5B-Instruct/commit/b9d14db30f3849a06f43e2c7d5d413223919f5ed","source":"backfill"},{"id":"Qwen/Qwen2.5-0.5B-Instruct","date":"2024-09-18T01:38:51.000Z","classes":["template"],"sentence":"Qwen2.5-0.5B-Instruct: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2024-09-18).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"89d726e7795b1019","after":"d5495a1e5db06111"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"eaa56b503cc0a8a4d15de1dd8bd2a7e95a716be2","toSha":"a338b55dd21219a5f4da42bc11a9313d1a27d4cc","commit":"https://huggingface.co/Qwen/Qwen2.5-0.5B-Instruct/commit/a338b55dd21219a5f4da42bc11a9313d1a27d4cc","source":"backfill"},{"id":"Qwen/Qwen2.5-72B-Instruct","date":"2024-09-17T19:04:31.000Z","classes":["template"],"sentence":"Qwen2.5-72B-Instruct: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2024-09-17).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"89d726e7795b1019","after":"d5495a1e5db06111"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"d9c31670291b860cecf111a94ce34c4e3ab66fcc","toSha":"f073433cb484002b27d7a84e8bce1c7435e14a1c","commit":"https://huggingface.co/Qwen/Qwen2.5-72B-Instruct/commit/f073433cb484002b27d7a84e8bce1c7435e14a1c","source":"backfill"},{"id":"Qwen/Qwen2.5-Coder-7B","date":"2024-09-17T14:35:19.000Z","classes":["structural"],"sentence":"Qwen2.5-Coder-7B: rope_scaling added ({\"factor\":4,\"original_max_position_embeddings\":32768,\"typ...); positions are now scaled (2024-09-17).","keys":[{"file":"config","key":"rope_scaling","before":null,"after":{"factor":4,"original_max_position_embeddings":32768,"type":"yarn"}}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"e1a210cef2e75b4f3d438e32efe3363b686641c7","toSha":"f52e93b2dd070f1c38fd0c403634261052d5c0fb","commit":"https://huggingface.co/Qwen/Qwen2.5-Coder-7B/commit/f52e93b2dd070f1c38fd0c403634261052d5c0fb","source":"backfill"},{"id":"Qwen/Qwen2.5-7B-Instruct","date":"2024-09-17T14:33:15.000Z","classes":["structural"],"sentence":"Qwen2.5-7B-Instruct: sliding_window 32768 to 131072 (2024-09-17).","keys":[{"file":"config","key":"sliding_window","before":32768,"after":131072}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"d1e200fcf95ef0d4326873ddf63e5562d5f1fdbb","toSha":"605f1116cfdfd9f5d5eb7d3f9ee2c546d1b1a0e3","commit":"https://huggingface.co/Qwen/Qwen2.5-7B-Instruct/commit/605f1116cfdfd9f5d5eb7d3f9ee2c546d1b1a0e3","source":"backfill"},{"id":"Qwen/Qwen2.5-Math-1.5B","date":"2024-09-17T12:16:35.000Z","classes":["template"],"sentence":"Qwen2.5-Math-1.5B: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2024-09-17).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"89d726e7795b1019","after":"4302efe46d5ecfe0"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"b53e4e3adbbb22508f16fbcdd6d0583175d6863b","toSha":"37193e87a152563370b983808fa02a8b7d1281bf","commit":"https://huggingface.co/Qwen/Qwen2.5-Math-1.5B/commit/37193e87a152563370b983808fa02a8b7d1281bf","source":"backfill"},{"id":"Qwen/Qwen2.5-72B-Instruct","date":"2024-09-16T20:16:47.000Z","classes":["structural"],"sentence":"Qwen2.5-72B-Instruct: sliding_window 32768 to 131072 (2024-09-16).","keys":[{"file":"config","key":"sliding_window","before":32768,"after":131072}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"f689aa03cd0893b424bb44b7292ca01f11dc6626","toSha":"d9c31670291b860cecf111a94ce34c4e3ab66fcc","commit":"https://huggingface.co/Qwen/Qwen2.5-72B-Instruct/commit/d9c31670291b860cecf111a94ce34c4e3ab66fcc","source":"backfill"},{"id":"Qwen/Qwen2.5-7B","date":"2024-09-16T11:28:55.000Z","classes":["template"],"sentence":"Qwen2.5-7B: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2024-09-16).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"793202280c0910ab","after":"89d726e7795b1019"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"7c128dfad5b6947573d3c14d6d8c423ed5be7af4","toSha":"3115c99c189e698dd7274e8a463a1025b0c878a6","commit":"https://huggingface.co/Qwen/Qwen2.5-7B/commit/3115c99c189e698dd7274e8a463a1025b0c878a6","source":"backfill"},{"id":"Qwen/Qwen2.5-1.5B","date":"2024-09-16T11:24:54.000Z","classes":["template"],"sentence":"Qwen2.5-1.5B: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2024-09-16).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"793202280c0910ab","after":"89d726e7795b1019"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"09e6328424220dbdff94a72e36498803ff47aa9e","toSha":"033b7782da89bcc457997494478bba1bf19025bc","commit":"https://huggingface.co/Qwen/Qwen2.5-1.5B/commit/033b7782da89bcc457997494478bba1bf19025bc","source":"backfill"},{"id":"Qwen/Qwen2.5-0.5B","date":"2024-09-16T11:23:53.000Z","classes":["template"],"sentence":"Qwen2.5-0.5B: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2024-09-16).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"793202280c0910ab","after":"89d726e7795b1019"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"b937cc073b55e980e6f21d5f8b7a77818efcf0eb","toSha":"74e0d9cfa1ba6f779d35f63aeb6cea7511937269","commit":"https://huggingface.co/Qwen/Qwen2.5-0.5B/commit/74e0d9cfa1ba6f779d35f63aeb6cea7511937269","source":"backfill"},{"id":"NousResearch/Hermes-3-Llama-3.1-8B","date":"2024-09-08T07:39:55.000Z","classes":["template"],"sentence":"Hermes-3-Llama-3.1-8B: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2024-09-08).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"e8d5f1a2a25432bb","after":"4b46d39895608509"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"489839286ef3415bd8c689dcdf96f812291e81ba","toSha":"896ea440e5a9e6070e3d8a2774daf2b481ab425b","commit":"https://huggingface.co/NousResearch/Hermes-3-Llama-3.1-8B/commit/896ea440e5a9e6070e3d8a2774daf2b481ab425b","source":"backfill"},{"id":"NousResearch/Hermes-3-Llama-3.1-8B","date":"2024-09-08T07:25:22.000Z","classes":["tokenizer"],"sentence":"Hermes-3-Llama-3.1-8B: tokenizer_config.json was removed; whatever defaults it set now come from the library (2024-09-08).","keys":[{"file":"tokenizer","key":"(file)","before":"present","after":null}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"5ba4922000a4050f66d359446248817b40479c37","toSha":"ebcf7c1aca87a99f70e26cf59678587dd73ed4bf","commit":"https://huggingface.co/NousResearch/Hermes-3-Llama-3.1-8B/commit/ebcf7c1aca87a99f70e26cf59678587dd73ed4bf","source":"backfill"},{"id":"NousResearch/Hermes-3-Llama-3.1-8B","date":"2024-09-08T06:46:52.000Z","classes":["template"],"sentence":"Hermes-3-Llama-3.1-8B: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2024-09-08).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"34b01bc48aeb4380","after":"444594358972a4f6"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"5e3fd4817e213a9fd15a5bb3ec95788a6f04d4a5","toSha":"5ba4922000a4050f66d359446248817b40479c37","commit":"https://huggingface.co/NousResearch/Hermes-3-Llama-3.1-8B/commit/5ba4922000a4050f66d359446248817b40479c37","source":"backfill"},{"id":"NousResearch/Hermes-3-Llama-3.1-8B","date":"2024-09-05T07:35:38.000Z","classes":["template"],"sentence":"Hermes-3-Llama-3.1-8B: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2024-09-05).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"d6a8d8d9dd463870","after":"34b01bc48aeb4380"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"aa50cfc2048ce84e9734f79eb1cd76f8ed543661","toSha":"5e3fd4817e213a9fd15a5bb3ec95788a6f04d4a5","commit":"https://huggingface.co/NousResearch/Hermes-3-Llama-3.1-8B/commit/5e3fd4817e213a9fd15a5bb3ec95788a6f04d4a5","source":"backfill"},{"id":"nomic-ai/nomic-embed-text-v1","date":"2024-08-26T20:51:44.000Z","classes":["structural"],"sentence":"nomic-embed-text-v1: declared dtype \"bfloat16\" to \"float32\" (2024-08-26).","keys":[{"file":"config","key":"dtype","before":"bfloat16","after":"float32"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"282e906bdee9a76b26826a33bb976cfc8db78eaf","toSha":"cc62377b015c53a3bf52bb2f4eb8c55326d3f162","commit":"https://huggingface.co/nomic-ai/nomic-embed-text-v1/commit/cc62377b015c53a3bf52bb2f4eb8c55326d3f162","source":"backfill"},{"id":"nomic-ai/nomic-embed-text-v1.5","date":"2024-08-26T20:51:30.000Z","classes":["structural"],"sentence":"nomic-embed-text-v1.5: declared dtype \"bfloat16\" to \"float32\" (2024-08-26).","keys":[{"file":"config","key":"dtype","before":"bfloat16","after":"float32"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"c810decee00ecca9d3aca2e355cd16d7c85131dc","toSha":"679199c2575b5bfe93b06161d06cd7c16ebe4124","commit":"https://huggingface.co/nomic-ai/nomic-embed-text-v1.5/commit/679199c2575b5bfe93b06161d06cd7c16ebe4124","source":"backfill"},{"id":"nomic-ai/nomic-embed-text-v1","date":"2024-08-26T20:48:34.000Z","classes":["structural"],"sentence":"nomic-embed-text-v1: declared dtype \"float32\" to \"bfloat16\" (2024-08-26).","keys":[{"file":"config","key":"dtype","before":"float32","after":"bfloat16"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"e0830beb3bdd2a163d16d49c47f313d8659778d0","toSha":"282e906bdee9a76b26826a33bb976cfc8db78eaf","commit":"https://huggingface.co/nomic-ai/nomic-embed-text-v1/commit/282e906bdee9a76b26826a33bb976cfc8db78eaf","source":"backfill"},{"id":"nomic-ai/nomic-embed-text-v1.5","date":"2024-08-26T20:47:49.000Z","classes":["structural"],"sentence":"nomic-embed-text-v1.5: declared dtype \"float32\" to \"bfloat16\" (2024-08-26).","keys":[{"file":"config","key":"dtype","before":"float32","after":"bfloat16"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"91d2d6bfdddf0b0da840f901b533e99bae30d757","toSha":"c810decee00ecca9d3aca2e355cd16d7c85131dc","commit":"https://huggingface.co/nomic-ai/nomic-embed-text-v1.5/commit/c810decee00ecca9d3aca2e355cd16d7c85131dc","source":"backfill"},{"id":"OpenGVLab/InternVL2-26B","date":"2024-08-22T18:51:22.000Z","classes":["tokenizer"],"sentence":"InternVL2-26B: EOS declared as [92542,92543] (was unset); generation that relied on the library default now stops on this token (2024-08-22).","keys":[{"file":"generation","key":"eos_token_id","before":null,"after":[92542,92543]}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":{"before":null,"after":[92542,92543]},"bos":null},"fromSha":"81ae45afc57ae386c6e3d54b074512167751b732","toSha":"77238d470e24756249771843b801191788ffdcde","commit":"https://huggingface.co/OpenGVLab/InternVL2-26B/commit/77238d470e24756249771843b801191788ffdcde","source":"backfill"},{"id":"OpenGVLab/InternVL2-2B","date":"2024-08-22T18:50:09.000Z","classes":["tokenizer"],"sentence":"InternVL2-2B: EOS declared as [92542,92543] (was unset); generation that relied on the library default now stops on this token (2024-08-22).","keys":[{"file":"generation","key":"eos_token_id","before":null,"after":[92542,92543]}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":{"before":null,"after":[92542,92543]},"bos":null},"fromSha":"19371f127fd344f84ba16c482a1f98c653b1f4cf","toSha":"fa2d59d853efd9ed57940c50dc9e7ce11e7cb4f6","commit":"https://huggingface.co/OpenGVLab/InternVL2-2B/commit/fa2d59d853efd9ed57940c50dc9e7ce11e7cb4f6","source":"backfill"},{"id":"OpenGVLab/InternVL2-1B","date":"2024-08-22T18:50:02.000Z","classes":["tokenizer"],"sentence":"InternVL2-1B: EOS declared as [151644,151645] (was unset); generation that relied on the library default now stops on this token (2024-08-22).","keys":[{"file":"generation","key":"eos_token_id","before":null,"after":[151644,151645]}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":{"before":null,"after":[151644,151645]},"bos":null},"fromSha":"055968f34f4258bd06fe751a2c8eda46e208d8e6","toSha":"8d827904f96b944be1d03fccd82e51a000c10d1d","commit":"https://huggingface.co/OpenGVLab/InternVL2-1B/commit/8d827904f96b944be1d03fccd82e51a000c10d1d","source":"backfill"},{"id":"NousResearch/Hermes-3-Llama-3.1-8B","date":"2024-08-22T10:12:10.000Z","classes":["template"],"sentence":"Hermes-3-Llama-3.1-8B: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2024-08-22).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"6d809ba2bea6b789","after":"d6a8d8d9dd463870"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"67bf4aca9f4243e275f402f3708eed3aa8a9038c","toSha":"aa50cfc2048ce84e9734f79eb1cd76f8ed543661","commit":"https://huggingface.co/NousResearch/Hermes-3-Llama-3.1-8B/commit/aa50cfc2048ce84e9734f79eb1cd76f8ed543661","source":"backfill"},{"id":"NousResearch/Hermes-3-Llama-3.1-8B","date":"2024-08-18T22:00:20.000Z","classes":["template"],"sentence":"Hermes-3-Llama-3.1-8B: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2024-08-18).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"a805e50fed68938a","after":"6d809ba2bea6b789"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"db6a44dd23ecd3eb23b46234d44214998ab38540","toSha":"f537e1048542f59a6ac3d2662ff36dd756755dde","commit":"https://huggingface.co/NousResearch/Hermes-3-Llama-3.1-8B/commit/f537e1048542f59a6ac3d2662ff36dd756755dde","source":"backfill"},{"id":"meta-llama/Llama-3.1-405B-FP8","date":"2024-08-14T13:27:04.000Z","classes":["structural"],"sentence":"Llama-3.1-405B-FP8: KV heads 16 to 8; derived parameters 410.08B to 405.85B (-4.23B, -8.46 GB of bf16 weights) (2024-08-14).","keys":[{"file":"config","key":"num_key_value_heads","before":16,"after":8}],"consequence":{"params":{"before":410081375488,"after":405853517056},"weightsGbBf16":{"before":820.163,"after":811.707},"trainPeakGb":{"before":12982.474,"after":12855.976},"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"60db52617fa6b1d8ad96b37f6e6bcc172b6f6846","toSha":"d79e67b5ee26735510cff4fb9ec10fb74fadb8c1","commit":"https://huggingface.co/meta-llama/Llama-3.1-405B-FP8/commit/d79e67b5ee26735510cff4fb9ec10fb74fadb8c1","source":"backfill"},{"id":"unsloth/mistral-7b-v0.3-bnb-4bit","date":"2024-08-07T01:38:27.000Z","classes":["tokenizer"],"sentence":"mistral-7b-v0.3-bnb-4bit: pad token unset to 770; batched training masks a different id (2024-08-07).","keys":[{"file":"config","key":"pad_token_id","before":null,"after":770}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"ff1aba4d92518713a7366d03d247afb1b0b76a64","toSha":"51e07b480c7f568002ee69cbecc22f1518d016e3","commit":"https://huggingface.co/unsloth/mistral-7b-v0.3-bnb-4bit/commit/51e07b480c7f568002ee69cbecc22f1518d016e3","source":"backfill"},{"id":"unsloth/mistral-7b-v0.3-bnb-4bit","date":"2024-08-07T00:55:23.000Z","classes":["tokenizer","generation-defaults"],"sentence":"mistral-7b-v0.3-bnb-4bit: pad token unset to 770; batched training masks a different id; generation defaults changed (max_length unset to 32768); a call that does not pass these explicitly now generates differently (2024-08-07).","keys":[{"file":"generation","key":"max_length","before":null,"after":32768},{"file":"generation","key":"pad_token_id","before":null,"after":770}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"6ae0f3056d3ba783295883026e417e8f838a6e9e","toSha":"ff1aba4d92518713a7366d03d247afb1b0b76a64","commit":"https://huggingface.co/unsloth/mistral-7b-v0.3-bnb-4bit/commit/ff1aba4d92518713a7366d03d247afb1b0b76a64","source":"backfill"},{"id":"unsloth/mistral-7b-v0.3-bnb-4bit","date":"2024-08-07T00:55:22.000Z","classes":["structural","tokenizer"],"sentence":"mistral-7b-v0.3-bnb-4bit: head dim unset to 128; pad token 770 to unset; batched training masks a different id (2024-08-07).","keys":[{"file":"config","key":"head_dim","before":null,"after":128},{"file":"config","key":"pad_token_id","before":770,"after":null}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"fdeda262808c80a14999cf57fc891418b0569f8e","toSha":"6ae0f3056d3ba783295883026e417e8f838a6e9e","commit":"https://huggingface.co/unsloth/mistral-7b-v0.3-bnb-4bit/commit/6ae0f3056d3ba783295883026e417e8f838a6e9e","source":"backfill"},{"id":"unsloth/Meta-Llama-3.1-8B-Instruct","date":"2024-08-07T00:13:08.000Z","classes":["generation-defaults"],"sentence":"Meta-Llama-3.1-8B-Instruct: generation defaults changed (max_length unset to 131072); a call that does not pass these explicitly now generates differently (2024-08-07).","keys":[{"file":"generation","key":"max_length","before":null,"after":131072}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"e1ea12ba7c95ff18fc1bc5c558824facc6c70b2d","toSha":"e4df7a7b9c960325f60a186668edfda6d1369c5b","commit":"https://huggingface.co/unsloth/Meta-Llama-3.1-8B-Instruct/commit/e4df7a7b9c960325f60a186668edfda6d1369c5b","source":"backfill"},{"id":"unsloth/Meta-Llama-3.1-8B-Instruct","date":"2024-08-06T23:42:18.000Z","classes":["tokenizer"],"sentence":"Meta-Llama-3.1-8B-Instruct: pad token unset to 128004; batched training masks a different id (2024-08-06).","keys":[{"file":"config","key":"pad_token_id","before":null,"after":128004}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"cf3c8428e6ee045e37fe38e255ede0b043b133d9","toSha":"f9a49112db6e5913caa8170ac3a4725174b20f0c","commit":"https://huggingface.co/unsloth/Meta-Llama-3.1-8B-Instruct/commit/f9a49112db6e5913caa8170ac3a4725174b20f0c","source":"backfill"},{"id":"unsloth/Meta-Llama-3.1-8B-Instruct","date":"2024-08-06T23:42:18.000Z","classes":["tokenizer"],"sentence":"Meta-Llama-3.1-8B-Instruct: pad token unset to 128004; batched training masks a different id (2024-08-06).","keys":[{"file":"generation","key":"pad_token_id","before":null,"after":128004}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"f9a49112db6e5913caa8170ac3a4725174b20f0c","toSha":"e1ea12ba7c95ff18fc1bc5c558824facc6c70b2d","commit":"https://huggingface.co/unsloth/Meta-Llama-3.1-8B-Instruct/commit/e1ea12ba7c95ff18fc1bc5c558824facc6c70b2d","source":"backfill"},{"id":"unsloth/Meta-Llama-3.1-8B-Instruct","date":"2024-08-06T18:17:58.000Z","classes":["template"],"sentence":"Meta-Llama-3.1-8B-Instruct: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2024-08-06).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"b48c47f644389271","after":"e10ca381b1ccc5cf"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"a95354754cf0ac6345ac858cf9b56a719c6d33a4","toSha":"ec7283d262156dd5d1781224d6953903bc0ece2d","commit":"https://huggingface.co/unsloth/Meta-Llama-3.1-8B-Instruct/commit/ec7283d262156dd5d1781224d6953903bc0ece2d","source":"backfill"},{"id":"NousResearch/Hermes-3-Llama-3.1-8B","date":"2024-08-06T03:33:28.000Z","classes":["template"],"sentence":"Hermes-3-Llama-3.1-8B: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2024-08-06).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"58c1a1f04baa7ada","after":"a805e50fed68938a"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"248bb9b3e1f00f5e7d530090a17c9201eca42e09","toSha":"db6a44dd23ecd3eb23b46234d44214998ab38540","commit":"https://huggingface.co/NousResearch/Hermes-3-Llama-3.1-8B/commit/db6a44dd23ecd3eb23b46234d44214998ab38540","source":"backfill"},{"id":"NousResearch/Hermes-3-Llama-3.1-8B","date":"2024-07-31T01:27:19.000Z","classes":["tokenizer"],"sentence":"Hermes-3-Llama-3.1-8B: EOS changed from \"<|end_of_text|>\" to \"<|im_end|>\"; anything fine-tuned or served against the old value stops at a different token (2024-07-31).","keys":[{"file":"tokenizer","key":"eos_token","before":"<|end_of_text|>","after":"<|im_end|>"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":{"before":"<|end_of_text|>","after":"<|im_end|>"},"bos":null},"fromSha":"5aada9f97327e253b3cf836e051df623a965a162","toSha":"9949f486fbfa5834405a9aabf4adb26fe5b29274","commit":"https://huggingface.co/NousResearch/Hermes-3-Llama-3.1-8B/commit/9949f486fbfa5834405a9aabf4adb26fe5b29274","source":"backfill"},{"id":"NousResearch/Hermes-3-Llama-3.1-8B","date":"2024-07-31T01:26:06.000Z","classes":["tokenizer"],"sentence":"Hermes-3-Llama-3.1-8B: pad token \"<|end_of_text|>\" to \"<|im_end|>\"; batched training masks a different id (2024-07-31).","keys":[{"file":"tokenizer","key":"pad_token","before":"<|end_of_text|>","after":"<|im_end|>"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"34ccd9045c44ce074650c47bdb86254f1aa05956","toSha":"5aada9f97327e253b3cf836e051df623a965a162","commit":"https://huggingface.co/NousResearch/Hermes-3-Llama-3.1-8B/commit/5aada9f97327e253b3cf836e051df623a965a162","source":"backfill"},{"id":"NousResearch/Hermes-3-Llama-3.1-8B","date":"2024-07-31T01:19:41.000Z","classes":["tokenizer"],"sentence":"Hermes-3-Llama-3.1-8B: EOS changed from 128001 to 128040; anything fine-tuned or served against the old value stops at a different token (2024-07-31).","keys":[{"file":"config","key":"eos_token_id","before":128001,"after":128040}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":{"before":128001,"after":128040},"bos":null},"fromSha":"d4ed47f114cf92e2f6ad1eaa5c28dd7932ae4b64","toSha":"34ccd9045c44ce074650c47bdb86254f1aa05956","commit":"https://huggingface.co/NousResearch/Hermes-3-Llama-3.1-8B/commit/34ccd9045c44ce074650c47bdb86254f1aa05956","source":"backfill"},{"id":"NousResearch/Hermes-3-Llama-3.1-8B","date":"2024-07-31T01:19:29.000Z","classes":["tokenizer"],"sentence":"Hermes-3-Llama-3.1-8B: EOS changed from 128001 to 128040; anything fine-tuned or served against the old value stops at a different token (2024-07-31).","keys":[{"file":"generation","key":"eos_token_id","before":128001,"after":128040}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":{"before":128001,"after":128040},"bos":null},"fromSha":"46ed7f46fd64dd4b09cd0d663f71bb2b1177505d","toSha":"d4ed47f114cf92e2f6ad1eaa5c28dd7932ae4b64","commit":"https://huggingface.co/NousResearch/Hermes-3-Llama-3.1-8B/commit/d4ed47f114cf92e2f6ad1eaa5c28dd7932ae4b64","source":"backfill"},{"id":"meta-llama/Llama-3.1-8B-Instruct","date":"2024-07-30T12:46:01.000Z","classes":["template"],"sentence":"Llama-3.1-8B-Instruct: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2024-07-30).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"571eddf08f7eb2ae","after":"e10ca381b1ccc5cf"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"b2a4d0f33b41fcd59a6d31662cc63b8d53367e1e","toSha":"8c22764a7e3675c50d4c7c9a4edb474456022b16","commit":"https://huggingface.co/meta-llama/Llama-3.1-8B-Instruct/commit/8c22764a7e3675c50d4c7c9a4edb474456022b16","source":"backfill"},{"id":"meta-llama/Llama-3.1-8B-Instruct","date":"2024-07-29T15:53:07.000Z","classes":["template"],"sentence":"Llama-3.1-8B-Instruct: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2024-07-29).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"b48c47f644389271","after":"571eddf08f7eb2ae"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"4281e96c7cf5ab6b312ef0cb78373efa3976a9dd","toSha":"b2a4d0f33b41fcd59a6d31662cc63b8d53367e1e","commit":"https://huggingface.co/meta-llama/Llama-3.1-8B-Instruct/commit/b2a4d0f33b41fcd59a6d31662cc63b8d53367e1e","source":"backfill"},{"id":"google/gemma-2-2b-it","date":"2024-07-28T22:17:49.000Z","classes":["tokenizer"],"sentence":"gemma-2-2b-it: EOS changed from 1 to [1,107]; anything fine-tuned or served against the old value stops at a different token (2024-07-28).","keys":[{"file":"config","key":"eos_token_id","before":1,"after":[1,107]},{"file":"generation","key":"eos_token_id","before":1,"after":[1,107]}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":{"before":1,"after":[1,107]},"bos":null},"fromSha":"980ca64487769ce33dea6add2cac6e2f075acea4","toSha":"17a01657f5c87135bcdd0ec7abb4b2dece04408a","commit":"https://huggingface.co/google/gemma-2-2b-it/commit/17a01657f5c87135bcdd0ec7abb4b2dece04408a","source":"backfill"},{"id":"meta-llama/Llama-3.1-405B-FP8","date":"2024-07-24T06:18:07.000Z","classes":["structural"],"sentence":"Llama-3.1-405B-FP8: quantization_config changed; the stored format of the weights is different (2024-07-24).","keys":[{"file":"config","key":"quantization_config","before":{"activation_scale_ub":1200,"quant_method":"fbgemm_fp8","_rest":"b3a3f8a22f72af8c"},"after":{"activation_scale_ub":1200,"quant_method":"fbgemm_fp8","_rest":"bcfb50e1048b00db"}}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"b46ddc84f9bbe25409feac525b16819a05719a02","toSha":"60db52617fa6b1d8ad96b37f6e6bcc172b6f6846","commit":"https://huggingface.co/meta-llama/Llama-3.1-405B-FP8/commit/60db52617fa6b1d8ad96b37f6e6bcc172b6f6846","source":"backfill"},{"id":"meta-llama/Llama-3.1-405B-FP8","date":"2024-07-23T15:43:12.000Z","classes":["structural"],"sentence":"Llama-3.1-405B-FP8: rope_scaling added ({\"factor\":8,\"high_freq_factor\":4,\"low_freq_factor\":1,\"ori...); positions are now scaled (2024-07-23).","keys":[{"file":"config","key":"rope_scaling","before":null,"after":{"factor":8,"low_freq_factor":1,"high_freq_factor":4,"original_max_position_embeddings":8192,"rope_type":"llama3"}}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"f2528c44b8a305634d2e0586ad75e7fba673a6da","toSha":"b46ddc84f9bbe25409feac525b16819a05719a02","commit":"https://huggingface.co/meta-llama/Llama-3.1-405B-FP8/commit/b46ddc84f9bbe25409feac525b16819a05719a02","source":"backfill"},{"id":"meta-llama/Llama-3.1-8B","date":"2024-07-23T15:39:30.000Z","classes":["structural"],"sentence":"Llama-3.1-8B: rope_scaling added ({\"factor\":8,\"high_freq_factor\":4,\"low_freq_factor\":1,\"ori...); positions are now scaled (2024-07-23).","keys":[{"file":"config","key":"rope_scaling","before":null,"after":{"factor":8,"low_freq_factor":1,"high_freq_factor":4,"original_max_position_embeddings":8192,"rope_type":"llama3"}}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"9cb1f4b8e2ac981fe14a4df35165360c7dcb5aa3","toSha":"13f04ed6f85ef2aa2fd11b960a275c3e31a8069e","commit":"https://huggingface.co/meta-llama/Llama-3.1-8B/commit/13f04ed6f85ef2aa2fd11b960a275c3e31a8069e","source":"backfill"},{"id":"meta-llama/Llama-3.1-8B-Instruct","date":"2024-07-23T14:51:30.000Z","classes":["structural"],"sentence":"Llama-3.1-8B-Instruct: rope_scaling added ({\"factor\":8,\"high_freq_factor\":4,\"low_freq_factor\":1,\"ori...); positions are now scaled (2024-07-23).","keys":[{"file":"config","key":"rope_scaling","before":null,"after":{"factor":8,"low_freq_factor":1,"high_freq_factor":4,"original_max_position_embeddings":8192,"rope_type":"llama3"}}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"6d52333029e2eee3b787f7457bbfb4c725f7787d","toSha":"91e702621d649240f67086276478556294b22eeb","commit":"https://huggingface.co/meta-llama/Llama-3.1-8B-Instruct/commit/91e702621d649240f67086276478556294b22eeb","source":"backfill"},{"id":"meta-llama/Llama-3.1-8B","date":"2024-07-23T07:13:23.000Z","classes":["generation-defaults"],"sentence":"Llama-3.1-8B: generation defaults changed (do_sample unset to true, temperature unset to 0.6, top_p unset to 0.9); a call that does not pass these explicitly now generates differently (2024-07-23).","keys":[{"file":"generation","key":"do_sample","before":null,"after":true},{"file":"generation","key":"temperature","before":null,"after":0.6},{"file":"generation","key":"top_p","before":null,"after":0.9}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"b9701d0be774a86805582ec5b34798ec5a106aef","toSha":"9cb1f4b8e2ac981fe14a4df35165360c7dcb5aa3","commit":"https://huggingface.co/meta-llama/Llama-3.1-8B/commit/9cb1f4b8e2ac981fe14a4df35165360c7dcb5aa3","source":"backfill"},{"id":"meta-llama/Llama-3.1-405B-FP8","date":"2024-07-22T17:28:45.000Z","classes":["generation-defaults"],"sentence":"Llama-3.1-405B-FP8: generation defaults changed (do_sample unset to true, temperature unset to 0.6, top_p unset to 0.9); a call that does not pass these explicitly now generates differently (2024-07-22).","keys":[{"file":"generation","key":"do_sample","before":null,"after":true},{"file":"generation","key":"temperature","before":null,"after":0.6},{"file":"generation","key":"top_p","before":null,"after":0.9}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"1965e1c192712e46edf097e4c4dced4540510125","toSha":"f2528c44b8a305634d2e0586ad75e7fba673a6da","commit":"https://huggingface.co/meta-llama/Llama-3.1-405B-FP8/commit/f2528c44b8a305634d2e0586ad75e7fba673a6da","source":"backfill"},{"id":"meta-llama/Llama-3.1-405B-FP8","date":"2024-07-21T00:18:55.000Z","classes":["metadata"],"sentence":"Llama-3.1-405B-FP8: declared dtype unset to \"bfloat16\" (2024-07-21).","keys":[{"file":"config","key":"dtype","before":null,"after":"bfloat16"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"e981b1a797076be9c34e893c228f2c1cffc7f336","toSha":"79d637ba1eea1e7fabf7d20554b8330eb06912b3","commit":"https://huggingface.co/meta-llama/Llama-3.1-405B-FP8/commit/79d637ba1eea1e7fabf7d20554b8330eb06912b3","source":"backfill"},{"id":"microsoft/Florence-2-large","date":"2024-07-20T00:13:01.000Z","classes":["structural"],"sentence":"Florence-2-large: declared dtype \"float32\" to \"float16\" (2024-07-20).","keys":[{"file":"config","key":"dtype","before":"float32","after":"float16"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"8651614293449a83056b97d042339b17b2f3d14a","toSha":"6bf179230dd8855083a51a5e11beb04aec1291fd","commit":"https://huggingface.co/microsoft/Florence-2-large/commit/6bf179230dd8855083a51a5e11beb04aec1291fd","source":"backfill"},{"id":"microsoft/Florence-2-base","date":"2024-07-20T00:12:40.000Z","classes":["structural"],"sentence":"Florence-2-base: declared dtype \"float32\" to \"float16\" (2024-07-20).","keys":[{"file":"config","key":"dtype","before":"float32","after":"float16"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"78e1dd41f4c83f8c3e2c60c73057caa8a570186a","toSha":"ee1f1f163f352801f3b7af6b2b96e4baaa6ff2ff","commit":"https://huggingface.co/microsoft/Florence-2-base/commit/ee1f1f163f352801f3b7af6b2b96e4baaa6ff2ff","source":"backfill"},{"id":"microsoft/Florence-2-base-ft","date":"2024-07-20T00:12:30.000Z","classes":["structural"],"sentence":"Florence-2-base-ft: declared dtype \"float32\" to \"float16\" (2024-07-20).","keys":[{"file":"config","key":"dtype","before":"float32","after":"float16"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"9cef882064569cdc77847a7be06ce7fa3a9f2541","toSha":"9803f52844ec1ae5df004e6089262e9a23e527fd","commit":"https://huggingface.co/microsoft/Florence-2-base-ft/commit/9803f52844ec1ae5df004e6089262e9a23e527fd","source":"backfill"},{"id":"meta-llama/Llama-3.1-8B","date":"2024-07-18T18:57:49.000Z","classes":["tokenizer"],"sentence":"Llama-3.1-8B: tokenizer model_max_length 1e+30 to 131072; truncation defaults change (2024-07-18).","keys":[{"file":"tokenizer","key":"model_max_length","before":1e+30,"after":131072}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"e5c39e551424c763dbc3e58e32ef2999d33a6d8d","toSha":"b9701d0be774a86805582ec5b34798ec5a106aef","commit":"https://huggingface.co/meta-llama/Llama-3.1-8B/commit/b9701d0be774a86805582ec5b34798ec5a106aef","source":"backfill"},{"id":"meta-llama/Llama-3.1-8B-Instruct","date":"2024-07-18T18:11:08.000Z","classes":["tokenizer"],"sentence":"Llama-3.1-8B-Instruct: EOS changed from [128001,128009] to [128001,128008,128009]; anything fine-tuned or served against the old value stops at a different token (2024-07-18).","keys":[{"file":"config","key":"eos_token_id","before":[128001,128009],"after":[128001,128008,128009]}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":{"before":[128001,128009],"after":[128001,128008,128009]},"bos":null},"fromSha":"9b17df7da8b89e577045f3617f4af89796182461","toSha":"6d52333029e2eee3b787f7457bbfb4c725f7787d","commit":"https://huggingface.co/meta-llama/Llama-3.1-8B-Instruct/commit/6d52333029e2eee3b787f7457bbfb4c725f7787d","source":"backfill"},{"id":"meta-llama/Llama-3.1-8B-Instruct","date":"2024-07-18T18:08:01.000Z","classes":["tokenizer"],"sentence":"Llama-3.1-8B-Instruct: EOS changed from [128001,128009] to [128001,128008,128009]; anything fine-tuned or served against the old value stops at a different token (2024-07-18).","keys":[{"file":"generation","key":"eos_token_id","before":[128001,128009],"after":[128001,128008,128009]}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":{"before":[128001,128009],"after":[128001,128008,128009]},"bos":null},"fromSha":"bedbcc3b91b656331bf7569fa6661dad0cec360f","toSha":"9b17df7da8b89e577045f3617f4af89796182461","commit":"https://huggingface.co/meta-llama/Llama-3.1-8B-Instruct/commit/9b17df7da8b89e577045f3617f4af89796182461","source":"backfill"},{"id":"jinaai/jina-reranker-v2-base-multilingual","date":"2024-07-17T13:50:09.000Z","classes":["tokenizer"],"sentence":"jina-reranker-v2-base-multilingual: tokenizer model_max_length 1026 to 1024; truncation defaults change (2024-07-17).","keys":[{"file":"tokenizer","key":"model_max_length","before":1026,"after":1024}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"67b96c49dc51a00c7386c68d83e39ef0fd6a3331","toSha":"6371035883584acb0584c34dbaa9350e4752b8d2","commit":"https://huggingface.co/jinaai/jina-reranker-v2-base-multilingual/commit/6371035883584acb0584c34dbaa9350e4752b8d2","source":"backfill"},{"id":"meta-llama/Llama-3.1-8B","date":"2024-07-17T09:34:50.000Z","classes":["template"],"sentence":"Llama-3.1-8B: the chat template was removed (2024-07-17).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"b48c47f644389271","after":null}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"f6259a50a13f2bdc59b651fc272027fb8528f276","toSha":"e5c39e551424c763dbc3e58e32ef2999d33a6d8d","commit":"https://huggingface.co/meta-llama/Llama-3.1-8B/commit/e5c39e551424c763dbc3e58e32ef2999d33a6d8d","source":"backfill"},{"id":"meta-llama/Llama-3.1-8B","date":"2024-07-16T21:47:51.000Z","classes":["structural"],"sentence":"Llama-3.1-8B: context window 16384 to 131072 tokens (2024-07-16).","keys":[{"file":"config","key":"max_position_embeddings","before":16384,"after":131072}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":{"before":16384,"after":131072},"eos":null,"bos":null},"fromSha":"01d63cd05cf2f6b5151f87509c7a5954d8a64e62","toSha":"f6259a50a13f2bdc59b651fc272027fb8528f276","commit":"https://huggingface.co/meta-llama/Llama-3.1-8B/commit/f6259a50a13f2bdc59b651fc272027fb8528f276","source":"backfill"},{"id":"meta-llama/Llama-3.1-8B","date":"2024-07-16T15:50:35.000Z","classes":["structural"],"sentence":"Llama-3.1-8B: context window 8192 to 16384 tokens (2024-07-16).","keys":[{"file":"config","key":"max_position_embeddings","before":8192,"after":16384}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":{"before":8192,"after":16384},"eos":null,"bos":null},"fromSha":"63565f4742196ac840b9ff168f57cae073b2646b","toSha":"01d63cd05cf2f6b5151f87509c7a5954d8a64e62","commit":"https://huggingface.co/meta-llama/Llama-3.1-8B/commit/01d63cd05cf2f6b5151f87509c7a5954d8a64e62","source":"backfill"},{"id":"google/gemma-2-2b-it","date":"2024-07-16T11:31:21.000Z","classes":["template"],"sentence":"gemma-2-2b-it: a chat template was added; apply_chat_template now formats prompts that it previously refused or formatted by default (2024-07-16).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":null,"after":"ecd6ae513fe103f0"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"44317f247eebc0288d54496f1eb22754dac637c9","toSha":"d8d6c73026ed2e8c636a08d24d647bb427ab0325","commit":"https://huggingface.co/google/gemma-2-2b-it/commit/d8d6c73026ed2e8c636a08d24d647bb427ab0325","source":"backfill"},{"id":"mistralai/Mistral-7B-v0.1","date":"2024-07-04T14:35:47.000Z","classes":["template"],"sentence":"Mistral-7B-v0.1: the chat template was removed (2024-07-04).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"796853172235ab3a","after":null}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"67b958e1e004930a761765ead7c72133cdbfe68d","toSha":"f9824c3c1090baa2f7c1d33ddc89becab37b3e18","commit":"https://huggingface.co/mistralai/Mistral-7B-v0.1/commit/f9824c3c1090baa2f7c1d33ddc89becab37b3e18","source":"backfill"},{"id":"mistralai/Mistral-7B-v0.1","date":"2024-07-03T22:15:07.000Z","classes":["template"],"sentence":"Mistral-7B-v0.1: a chat template was added; apply_chat_template now formats prompts that it previously refused or formatted by default (2024-07-03).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":null,"after":"796853172235ab3a"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"17688039ea7b7001c702797eed2ab7716a0cc3c2","toSha":"67b958e1e004930a761765ead7c72133cdbfe68d","commit":"https://huggingface.co/mistralai/Mistral-7B-v0.1/commit/67b958e1e004930a761765ead7c72133cdbfe68d","source":"backfill"},{"id":"mistralai/Mistral-7B-Instruct-v0.2","date":"2024-07-03T22:10:26.000Z","classes":["template"],"sentence":"Mistral-7B-Instruct-v0.2: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2024-07-03).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"18f104df66d35dcc","after":"796853172235ab3a"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"1296dc8fd9b21e6424c9c305c06db9ae60c03ace","toSha":"250544c9a802b0396550d0fd24bc80ff98bb1f5f","commit":"https://huggingface.co/mistralai/Mistral-7B-Instruct-v0.2/commit/250544c9a802b0396550d0fd24bc80ff98bb1f5f","source":"backfill"},{"id":"OpenGVLab/InternVL2-2B","date":"2024-07-03T16:09:13.000Z","classes":["structural"],"sentence":"InternVL2-2B: rope_scaling added ({\"factor\":2,\"type\":\"dynamic\"}); positions are now scaled (2024-07-03).","keys":[{"file":"config","key":"rope_scaling","before":null,"after":{"factor":2,"type":"dynamic"}}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"02eb177c2c84a2f8dbd466cc40ec5ad52881a550","toSha":"9816b63a5c20c173eabc9a11003808f5b76b1532","commit":"https://huggingface.co/OpenGVLab/InternVL2-2B/commit/9816b63a5c20c173eabc9a11003808f5b76b1532","source":"backfill"},{"id":"microsoft/Phi-3-mini-128k-instruct","date":"2024-07-01T20:40:25.000Z","classes":["structural","tokenizer","template"],"sentence":"Phi-3-mini-128k-instruct: rope_scaling values changed ({\"long_factor\":[1.0299999713897705,1.0499999523162842,1.0... to {\"long_factor\":[1.0700000524520874,1.1200000047683716,1.1...); positions are computed differently; the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering; add_bos_token true to false; encoding the same text yields different ids (2024-07-01).","keys":[{"file":"config","key":"rope_scaling","before":{"long_factor":[1.0299999713897705,1.0499999523162842,1.0499999523162842,1.0799999237060547,1.2299998998641968,1.2299998998641968,1.2999999523162842,1.4499999284744263,1.5999999046325684,1.6499998569488525,1.8999998569488525,2.859999895095825,3.68999981880188,5.419999599456787,5.489999771118164,5.489999771118164,9.09000015258789,11.579999923706055,15.65999984741211,15.769999504089355,15.789999961853027,18.360000610351562,21.989999771118164,23.079999923706055,30.009998321533203,32.35000228881836,32.590003967285156,35.56000518798828,39.95000457763672,53.840003967285156,56.20000457763672,57.95000457763672,59.29000473022461,59.77000427246094,59.920005798339844,61.190006256103516,61.96000671386719,62.50000762939453,63.3700065612793,63.48000717163086,63.48000717163086,63.66000747680664,63.850006103515625,64.08000946044922,64.760009765625,64.80001068115234,64.81001281738281,64.81001281738281],"short_factor":[1.05,1.05,1.05,1.1,1.1,1.1500000000000001,1.2000000000000002,1.2500000000000002,1.3000000000000003,1.3500000000000003,1.5000000000000004,2.000000000000001,2.000000000000001,2.000000000000001,2.000000000000001,2.000000000000001,2.000000000000001,2.000000000000001,2.000000000000001,2.000000000000001,2.000000000000001,2.000000000000001,2.000000000000001,2.000000000000001,2.000000000000001,2.000000000000001,2.000000000000001,2.000000000000001,2.000000000000001,2.000000000000001,2.000000000000001,2.000000000000001,2.0500000000000007,2.0500000000000007,2.0500000000000007,2.1000000000000005,2.1000000000000005,2.1000000000000005,2.1500000000000004,2.1500000000000004,2.3499999999999996,2.549999999999999,2.5999999999999988,2.5999999999999988,2.7499999999999982,2.849999999999998,2.849999999999998,2.9499999999999975],"type":"su"},"after":{"long_factor":[1.0700000524520874,1.1200000047683716,1.149999976158142,1.4199999570846558,1.5699999332427979,1.7999999523162842,2.129999876022339,2.129999876022339,3.009999990463257,5.910000324249268,6.950000286102295,9.070000648498535,9.930000305175781,10.710000038146973,11.130000114440918,14.609999656677246,15.409998893737793,19.809999465942383,37.279998779296875,38.279998779296875,38.599998474121094,40.12000274658203,46.20000457763672,50.940006256103516,53.66000747680664,54.9373893737793,56.89738845825195,57.28738784790039,59.98738479614258,60.86738586425781,60.887386322021484,61.71739196777344,62.91739273071289,62.957393646240234,63.41739273071289,63.8173942565918,63.83739471435547,63.897396087646484,63.93739700317383,64.06739807128906,64.11434936523438,64.12435150146484,64.15435028076172,64.19435119628906,64.24435424804688,64.57435607910156,64.69000244140625,64.76000213623047],"short_factor":[1.1,1.1,1.1,1.3000000000000003,1.3500000000000003,1.3500000000000003,1.4000000000000004,1.5500000000000005,2.000000000000001,2.000000000000001,2.000000000000001,2.000000000000001,2.000000000000001,2.000000000000001,2.000000000000001,2.000000000000001,2.000000000000001,2.000000000000001,2.000000000000001,2.000000000000001,2.000000000000001,2.000000000000001,2.000000000000001,2.000000000000001,2.000000000000001,2.0500000000000007,2.0500000000000007,2.0500000000000007,2.0500000000000007,2.0500000000000007,2.0500000000000007,2.1000000000000005,2.1000000000000005,2.1500000000000004,2.25,2.25,2.25,2.25,2.25,2.3999999999999995,2.4499999999999993,2.499999999999999,2.6999999999999984,2.6999999999999984,2.7499999999999982,2.799999999999998,2.8999999999999977,3.049999999999997],"type":"longrope"}},{"file":"tokenizer","key":"add_bos_token","before":true,"after":false},{"file":"tokenizer","key":"chat_template_sha256","before":"268b6082ceb7176d","after":"dcaee66df77bfbb7"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"bb5bf1e4001277a606e11debca0ef80323e5f824","toSha":"10d25dfa1593265daf4a3bac573ab76ccf61d60f","commit":"https://huggingface.co/microsoft/Phi-3-mini-128k-instruct/commit/10d25dfa1593265daf4a3bac573ab76ccf61d60f","source":"backfill"},{"id":"microsoft/Phi-3-mini-4k-instruct","date":"2024-07-01T20:40:14.000Z","classes":["tokenizer","template"],"sentence":"Phi-3-mini-4k-instruct: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering; add_bos_token true to false; encoding the same text yields different ids (2024-07-01).","keys":[{"file":"tokenizer","key":"add_bos_token","before":true,"after":false},{"file":"tokenizer","key":"chat_template_sha256","before":"268b6082ceb7176d","after":"dcaee66df77bfbb7"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"ff07dc01615f8113924aed013115ab2abd32115b","toSha":"4f818b18e097c9ae8f93a29a57027cad54b75304","commit":"https://huggingface.co/microsoft/Phi-3-mini-4k-instruct/commit/4f818b18e097c9ae8f93a29a57027cad54b75304","source":"backfill"},{"id":"microsoft/Florence-2-base-ft","date":"2024-06-26T07:28:47.000Z","classes":["tokenizer"],"sentence":"Florence-2-base-ft: EOS changed from 1 to 2; anything fine-tuned or served against the old value stops at a different token; BOS changed from 2 to 0; prompts built by the old tokenizer start with a different token; pad token 0 to 1; batched training masks a different id (2024-06-26).","keys":[{"file":"config","key":"bos_token_id","before":2,"after":0},{"file":"config","key":"eos_token_id","before":1,"after":2},{"file":"config","key":"pad_token_id","before":0,"after":1}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":{"before":1,"after":2},"bos":{"before":2,"after":0}},"fromSha":"494b983c36978a56fdd3b1082f5995cc9fed7c09","toSha":"9cef882064569cdc77847a7be06ce7fa3a9f2541","commit":"https://huggingface.co/microsoft/Florence-2-base-ft/commit/9cef882064569cdc77847a7be06ce7fa3a9f2541","source":"backfill"},{"id":"microsoft/Florence-2-base","date":"2024-06-26T07:27:49.000Z","classes":["tokenizer"],"sentence":"Florence-2-base: EOS changed from 1 to 2; anything fine-tuned or served against the old value stops at a different token; BOS changed from 2 to 0; prompts built by the old tokenizer start with a different token; pad token 0 to 1; batched training masks a different id (2024-06-26).","keys":[{"file":"config","key":"bos_token_id","before":2,"after":0},{"file":"config","key":"eos_token_id","before":1,"after":2},{"file":"config","key":"pad_token_id","before":0,"after":1}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":{"before":1,"after":2},"bos":{"before":2,"after":0}},"fromSha":"a15cd1f8d0db4d1c41740e6754e22dba70bcfad6","toSha":"78e1dd41f4c83f8c3e2c60c73057caa8a570186a","commit":"https://huggingface.co/microsoft/Florence-2-base/commit/78e1dd41f4c83f8c3e2c60c73057caa8a570186a","source":"backfill"},{"id":"microsoft/Florence-2-large","date":"2024-06-26T07:23:53.000Z","classes":["tokenizer"],"sentence":"Florence-2-large: EOS changed from 1 to 2; anything fine-tuned or served against the old value stops at a different token; BOS changed from 2 to 0; prompts built by the old tokenizer start with a different token; pad token 0 to 1; batched training masks a different id (2024-06-26).","keys":[{"file":"config","key":"bos_token_id","before":2,"after":0},{"file":"config","key":"eos_token_id","before":1,"after":2},{"file":"config","key":"pad_token_id","before":0,"after":1}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":{"before":1,"after":2},"bos":{"before":2,"after":0}},"fromSha":"33ee8e109da7446c30db5ac65fa417853dd9cc9e","toSha":"8651614293449a83056b97d042339b17b2f3d14a","commit":"https://huggingface.co/microsoft/Florence-2-large/commit/8651614293449a83056b97d042339b17b2f3d14a","source":"backfill"},{"id":"mistralai/Mistral-7B-Instruct-v0.2","date":"2024-06-20T15:54:21.000Z","classes":["template"],"sentence":"Mistral-7B-Instruct-v0.2: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2024-06-20).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"26a59556925c9873","after":"18f104df66d35dcc"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"b70aa86578567ba3301b21c8a27bea4e8f6d6d61","toSha":"1296dc8fd9b21e6424c9c305c06db9ae60c03ace","commit":"https://huggingface.co/mistralai/Mistral-7B-Instruct-v0.2/commit/1296dc8fd9b21e6424c9c305c06db9ae60c03ace","source":"backfill"},{"id":"Qwen/Qwen2-1.5B-Instruct","date":"2024-06-06T09:10:54.000Z","classes":["generation-defaults"],"sentence":"Qwen2-1.5B-Instruct: generation defaults changed (temperature unset to 0.7, top_k unset to 20); a call that does not pass these explicitly now generates differently (2024-06-06).","keys":[{"file":"generation","key":"temperature","before":null,"after":0.7},{"file":"generation","key":"top_k","before":null,"after":20}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"8681046f92034373e55afbaddd7cf419f2c472d0","toSha":"1527f494deb8110df263e20032dbd96a11edc46a","commit":"https://huggingface.co/Qwen/Qwen2-1.5B-Instruct/commit/1527f494deb8110df263e20032dbd96a11edc46a","source":"backfill"},{"id":"Qwen/Qwen2-0.5B-Instruct","date":"2024-06-06T09:09:29.000Z","classes":["generation-defaults"],"sentence":"Qwen2-0.5B-Instruct: generation defaults changed (temperature unset to 0.7, top_k unset to 20); a call that does not pass these explicitly now generates differently (2024-06-06).","keys":[{"file":"generation","key":"temperature","before":null,"after":0.7},{"file":"generation","key":"top_k","before":null,"after":20}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"b17b4f1dd801064d41c30bc398e9435b5d0aa789","toSha":"bf61048b6a67ffd049ceab608702e7fc627852e5","commit":"https://huggingface.co/Qwen/Qwen2-0.5B-Instruct/commit/bf61048b6a67ffd049ceab608702e7fc627852e5","source":"backfill"},{"id":"jinaai/jina-clip-v1","date":"2024-06-03T09:36:34.000Z","classes":["structural"],"sentence":"jina-clip-v1: tie_word_embeddings true removed; the library default decides whether the head is shared (2024-06-03).","keys":[{"file":"config","key":"tie_word_embeddings","before":true,"after":null}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"1bae0621529ced998c73bca234a8cb9da997f33c","toSha":"a4f49a33555060b772543039ad3164cfbb693fc1","commit":"https://huggingface.co/jinaai/jina-clip-v1/commit/a4f49a33555060b772543039ad3164cfbb693fc1","source":"backfill"},{"id":"jinaai/jina-clip-v1","date":"2024-05-31T12:35:12.000Z","classes":["structural"],"sentence":"jina-clip-v1: config.json was removed; whatever defaults it set now come from the library (2024-05-31).","keys":[{"file":"config","key":"(file)","before":"present","after":null}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"84b04190c87ea500ec88c7815cf128b901799419","toSha":"0711c65ccee3d8fd96cf5af907a7c8387579dae1","commit":"https://huggingface.co/jinaai/jina-clip-v1/commit/0711c65ccee3d8fd96cf5af907a7c8387579dae1","source":"backfill"},{"id":"Qwen/Qwen2-0.5B","date":"2024-05-31T09:00:01.000Z","classes":["structural"],"sentence":"Qwen2-0.5B: sliding_window 4096 to 131072 (2024-05-31).","keys":[{"file":"config","key":"sliding_window","before":4096,"after":131072}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"b6aca0be68f0fe782c456140863b9e530a9a3f02","toSha":"ce45c397f0660f505c2885a6944478e8ee094b8f","commit":"https://huggingface.co/Qwen/Qwen2-0.5B/commit/ce45c397f0660f505c2885a6944478e8ee094b8f","source":"backfill"},{"id":"distilbert/distilbert-base-uncased","date":"2024-05-06T13:44:53.000Z","classes":["tokenizer"],"sentence":"distilbert-base-uncased: tokenizer model_max_length unset to 512; truncation defaults change (2024-05-06).","keys":[{"file":"tokenizer","key":"model_max_length","before":null,"after":512}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"b22dbc4a377fdc2c7115227abd6cc7fe55d4f49e","toSha":"12040accade4e8a0f71eabdb258fecc2e7e948be","commit":"https://huggingface.co/distilbert/distilbert-base-uncased/commit/12040accade4e8a0f71eabdb258fecc2e7e948be","source":"backfill"},{"id":"microsoft/Phi-3-mini-128k-instruct","date":"2024-04-29T11:56:06.000Z","classes":["template"],"sentence":"Phi-3-mini-128k-instruct: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2024-04-29).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"68e5b97472ceec75","after":"268b6082ceb7176d"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"48788d7704a9c3e2a7e44a3cb849deaaba701fa3","toSha":"fc313b0514906a348a624cc9421e36f52c60e137","commit":"https://huggingface.co/microsoft/Phi-3-mini-128k-instruct/commit/fc313b0514906a348a624cc9421e36f52c60e137","source":"backfill"},{"id":"microsoft/Phi-3-mini-4k-instruct","date":"2024-04-29T11:50:51.000Z","classes":["template"],"sentence":"Phi-3-mini-4k-instruct: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2024-04-29).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"68e5b97472ceec75","after":"268b6082ceb7176d"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"65be4e00a56c16d036e9cbe96b0b35f8aa0f84b0","toSha":"300945e90b6f55d3cb88261c8e5333fae696f672","commit":"https://huggingface.co/microsoft/Phi-3-mini-4k-instruct/commit/300945e90b6f55d3cb88261c8e5333fae696f672","source":"backfill"},{"id":"IlyaGusev/saiga_llama3_8b","date":"2024-04-26T06:20:18.000Z","classes":["tokenizer","template"],"sentence":"saiga_llama3_8b: EOS changed from \"<|im_end|>\" to \"<|eot_id|>\"; anything fine-tuned or served against the old value stops at a different token; BOS changed from \"<|im_start|>\" to \"<|begin_of_text|>\"; prompts built by the old tokenizer start with a different token; pad token \"<|im_start|>\" to \"<|begin_of_text|>\"; batched training masks a different id (2024-04-26).","keys":[{"file":"tokenizer","key":"bos_token","before":"<|im_start|>","after":"<|begin_of_text|>"},{"file":"tokenizer","key":"chat_template_sha256","before":"64cc1405a15b6b72","after":"ba03a121d097859c"},{"file":"tokenizer","key":"eos_token","before":"<|im_end|>","after":"<|eot_id|>"},{"file":"tokenizer","key":"pad_token","before":"<|im_start|>","after":"<|begin_of_text|>"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":{"before":"<|im_end|>","after":"<|eot_id|>"},"bos":{"before":"<|im_start|>","after":"<|begin_of_text|>"}},"fromSha":"91d7b8a51d0dde9658d6f715a01f1ec2cc2a9f50","toSha":"bc54c81cc63aa4836f64171f6fbb07d1f904b674","commit":"https://huggingface.co/IlyaGusev/saiga_llama3_8b/commit/bc54c81cc63aa4836f64171f6fbb07d1f904b674","source":"backfill"},{"id":"microsoft/Phi-3-mini-128k-instruct","date":"2024-04-25T18:48:51.000Z","classes":["tokenizer"],"sentence":"Phi-3-mini-128k-instruct: tokenizer model_max_length 4096 to 131072; truncation defaults change (2024-04-25).","keys":[{"file":"tokenizer","key":"model_max_length","before":4096,"after":131072}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"7de6f02aba40116fb61e55e27603a4c9b078e60c","toSha":"48788d7704a9c3e2a7e44a3cb849deaaba701fa3","commit":"https://huggingface.co/microsoft/Phi-3-mini-128k-instruct/commit/48788d7704a9c3e2a7e44a3cb849deaaba701fa3","source":"backfill"},{"id":"IlyaGusev/saiga_llama3_8b","date":"2024-04-25T18:24:52.000Z","classes":["tokenizer"],"sentence":"saiga_llama3_8b: EOS changed from 128001 to 128009; anything fine-tuned or served against the old value stops at a different token; pad token \"<|reserved_special_token_0|>\" to \"<|im_start|>\"; batched training masks a different id (2024-04-25).","keys":[{"file":"tokenizer","key":"pad_token","before":"<|reserved_special_token_0|>","after":"<|im_start|>"},{"file":"generation","key":"eos_token_id","before":128001,"after":128009},{"file":"generation","key":"pad_token_id","before":128002,"after":128000}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":{"before":128001,"after":128009},"bos":null},"fromSha":"d04705b3c527cc5f5b0ee48c170e4d921cf28d45","toSha":"91d7b8a51d0dde9658d6f715a01f1ec2cc2a9f50","commit":"https://huggingface.co/IlyaGusev/saiga_llama3_8b/commit/91d7b8a51d0dde9658d6f715a01f1ec2cc2a9f50","source":"backfill"},{"id":"IlyaGusev/saiga_llama3_8b","date":"2024-04-19T12:22:29.000Z","classes":["template"],"sentence":"saiga_llama3_8b: the chat template changed; prompts rendered with the old one no longer match what the model now expects, so cached prompts and eval sets need re-rendering (2024-04-19).","keys":[{"file":"tokenizer","key":"chat_template_sha256","before":"b48c47f644389271","after":"64cc1405a15b6b72"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"0af818c24e23e1aff7f3ad895c6862f4d3598971","toSha":"d04705b3c527cc5f5b0ee48c170e4d921cf28d45","commit":"https://huggingface.co/IlyaGusev/saiga_llama3_8b/commit/d04705b3c527cc5f5b0ee48c170e4d921cf28d45","source":"backfill"},{"id":"IlyaGusev/saiga_llama3_8b","date":"2024-04-19T11:52:31.000Z","classes":["tokenizer","generation-defaults"],"sentence":"saiga_llama3_8b: pad token unset to 128002; batched training masks a different id; generation defaults changed (do_sample unset to true, max_new_tokens unset to 1536, repetition_penalty unset to 1.12, temperature unset to 0.2); a call that does not pass these explicitly now generates differently (2024-04-19).","keys":[{"file":"generation","key":"do_sample","before":null,"after":true},{"file":"generation","key":"max_new_tokens","before":null,"after":1536},{"file":"generation","key":"pad_token_id","before":null,"after":128002},{"file":"generation","key":"repetition_penalty","before":null,"after":1.12},{"file":"generation","key":"temperature","before":null,"after":0.2},{"file":"generation","key":"top_k","before":null,"after":30},{"file":"generation","key":"top_p","before":null,"after":0.9}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"f4eb6c87c589f7df4da8618110df56f0d181bad4","toSha":"0af818c24e23e1aff7f3ad895c6862f4d3598971","commit":"https://huggingface.co/IlyaGusev/saiga_llama3_8b/commit/0af818c24e23e1aff7f3ad895c6862f4d3598971","source":"backfill"},{"id":"sentence-transformers/all-MiniLM-L12-v2","date":"2024-03-12T18:23:44.000Z","classes":["metadata"],"sentence":"all-MiniLM-L12-v2: architectures declared as [\"BertModel\"] (was unset) (2024-03-12).","keys":[{"file":"config","key":"architectures","before":null,"after":["BertModel"]}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"01e1a19f47cbdbe489da7bd30b8436620aeab787","toSha":"3392ab41e806eb6b71679072acbec53437ca20fd","commit":"https://huggingface.co/sentence-transformers/all-MiniLM-L12-v2/commit/3392ab41e806eb6b71679072acbec53437ca20fd","source":"backfill"},{"id":"google-bert/bert-base-uncased","date":"2024-02-19T11:06:12.000Z","classes":["tokenizer"],"sentence":"bert-base-uncased: tokenizer model_max_length unset to 512; truncation defaults change (2024-02-19).","keys":[{"file":"tokenizer","key":"model_max_length","before":null,"after":512}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"345fd30026bc3003828be943882dda32ab48b908","toSha":"86b5e0934494bd15c9632b12f734a8a67f723594","commit":"https://huggingface.co/google-bert/bert-base-uncased/commit/86b5e0934494bd15c9632b12f734a8a67f723594","source":"backfill"},{"id":"google-bert/bert-base-multilingual-uncased","date":"2024-02-19T11:06:00.000Z","classes":["tokenizer"],"sentence":"bert-base-multilingual-uncased: tokenizer model_max_length unset to 512; truncation defaults change (2024-02-19).","keys":[{"file":"tokenizer","key":"model_max_length","before":null,"after":512}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"d11aae01352cff41b1deab96882a87b9e4d13862","toSha":"7cbf9a625e29989f6b9c6c2fa68234c304f7e38f","commit":"https://huggingface.co/google-bert/bert-base-multilingual-uncased/commit/7cbf9a625e29989f6b9c6c2fa68234c304f7e38f","source":"backfill"},{"id":"google-bert/bert-base-cased","date":"2024-02-19T11:02:26.000Z","classes":["tokenizer"],"sentence":"bert-base-cased: tokenizer model_max_length unset to 512; truncation defaults change (2024-02-19).","keys":[{"file":"tokenizer","key":"model_max_length","before":null,"after":512}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"74d9687f45ab4451e3c5a87d61913de0e41b10f0","toSha":"cd5ef92a9fb2f889e972770a36d4ed042daf221e","commit":"https://huggingface.co/google-bert/bert-base-cased/commit/cd5ef92a9fb2f889e972770a36d4ed042daf221e","source":"backfill"},{"id":"jinaai/jina-embeddings-v2-base-code","date":"2024-02-05T07:13:55.000Z","classes":["structural","tokenizer"],"sentence":"jina-embeddings-v2-base-code: EOS declared as \"</s>\" (was unset); generation that relied on the library default now stops on this token; BOS declared as \"<s>\" (was unset); vocab_size changed from 30528 to 61056; the embedding and output head change shape, so adapters and merges made against the old vocabulary no longer load cleanly (2024-02-05).","keys":[{"file":"config","key":"dtype","before":"float32","after":"float16"},{"file":"config","key":"vocab_size","before":30528,"after":61056},{"file":"tokenizer","key":"bos_token","before":null,"after":"<s>"},{"file":"tokenizer","key":"eos_token","before":null,"after":"</s>"},{"file":"tokenizer","key":"model_max_length","before":2147483648,"after":8192},{"file":"tokenizer","key":"pad_token","before":"[PAD]","after":"<pad>"},{"file":"tokenizer","key":"tokenizer_class","before":"BertTokenizer","after":"RobertaTokenizer"},{"file":"tokenizer","key":"unk_token","before":"[UNK]","after":"<unk>"}],"consequence":{"params":{"before":114772992,"after":138218496},"weightsGbBf16":{"before":0.23,"after":0.276},"trainPeakGb":{"before":23.042,"after":23.492},"fingerprint":null,"context":null,"eos":{"before":null,"after":"</s>"},"bos":{"before":null,"after":"<s>"}},"fromSha":"a08d5882d94bfcc69188a4533e5c3bb1f197d37a","toSha":"a594cdd8ed6d6310f9dcada354356340e025cfb9","commit":"https://huggingface.co/jinaai/jina-embeddings-v2-base-code/commit/a594cdd8ed6d6310f9dcada354356340e025cfb9","source":"backfill"},{"id":"deepseek-ai/deepseek-coder-7b-instruct-v1.5","date":"2024-02-05T03:38:36.000Z","classes":["tokenizer"],"sentence":"deepseek-coder-7b-instruct-v1.5: EOS changed from 100001 to 100015; anything fine-tuned or served against the old value stops at a different token (2024-02-05).","keys":[{"file":"generation","key":"eos_token_id","before":100001,"after":100015}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":{"before":100001,"after":100015},"bos":null},"fromSha":"c63ca91ecafdb18a55364b095d752d3df1693c3f","toSha":"2a050a4c59d687a85324d32e147517992117ed30","commit":"https://huggingface.co/deepseek-ai/deepseek-coder-7b-instruct-v1.5/commit/2a050a4c59d687a85324d32e147517992117ed30","source":"backfill"},{"id":"deepseek-ai/deepseek-coder-7b-instruct-v1.5","date":"2024-02-05T03:38:18.000Z","classes":["tokenizer"],"sentence":"deepseek-coder-7b-instruct-v1.5: EOS changed from 100001 to 100015; anything fine-tuned or served against the old value stops at a different token (2024-02-05).","keys":[{"file":"config","key":"eos_token_id","before":100001,"after":100015}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":{"before":100001,"after":100015},"bos":null},"fromSha":"0c5229eb1c7a028c2f2fd51a0da070042ead8269","toSha":"c63ca91ecafdb18a55364b095d752d3df1693c3f","commit":"https://huggingface.co/deepseek-ai/deepseek-coder-7b-instruct-v1.5/commit/c63ca91ecafdb18a55364b095d752d3df1693c3f","source":"backfill"},{"id":"deepseek-ai/deepseek-coder-7b-instruct-v1.5","date":"2024-02-05T03:37:07.000Z","classes":["tokenizer"],"sentence":"deepseek-coder-7b-instruct-v1.5: EOS changed from 2 to 100001; anything fine-tuned or served against the old value stops at a different token; BOS changed from 1 to 100000; prompts built by the old tokenizer start with a different token (2024-02-05).","keys":[{"file":"generation","key":"bos_token_id","before":1,"after":100000},{"file":"generation","key":"eos_token_id","before":2,"after":100001}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":{"before":2,"after":100001},"bos":{"before":1,"after":100000}},"fromSha":"e66d398f7a1536339fbcac5f45d5828ffa5a337b","toSha":"0c5229eb1c7a028c2f2fd51a0da070042ead8269","commit":"https://huggingface.co/deepseek-ai/deepseek-coder-7b-instruct-v1.5/commit/0c5229eb1c7a028c2f2fd51a0da070042ead8269","source":"backfill"},{"id":"nomic-ai/nomic-embed-text-v1","date":"2024-02-01T14:58:15.000Z","classes":["structural"],"sentence":"nomic-embed-text-v1: context window 2048 to 8192 tokens (2024-02-01).","keys":[{"file":"config","key":"max_position_embeddings","before":2048,"after":8192}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":{"before":2048,"after":8192},"eos":null,"bos":null},"fromSha":"a350440d040317d78e36afb8441914d1f55f7126","toSha":"7952586070dd30d93aae9ffb1e405adf3bd1f352","commit":"https://huggingface.co/nomic-ai/nomic-embed-text-v1/commit/7952586070dd30d93aae9ffb1e405adf3bd1f352","source":"backfill"},{"id":"EleutherAI/gpt-neo-125m","date":"2024-01-31T20:29:39.000Z","classes":["tokenizer"],"sentence":"gpt-neo-125m: add_bos_token unset to false; encoding the same text yields different ids; tokenizer_class unset to \"GPT2Tokenizer\" (2024-01-31).","keys":[{"file":"tokenizer","key":"add_bos_token","before":null,"after":false},{"file":"tokenizer","key":"tokenizer_class","before":null,"after":"GPT2Tokenizer"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"0439444c18855fb88213e6e889bf87ffaa96c73e","toSha":"21def0189f5705e2521767faed922f1f15e7d7db","commit":"https://huggingface.co/EleutherAI/gpt-neo-125m/commit/21def0189f5705e2521767faed922f1f15e7d7db","source":"backfill"},{"id":"microsoft/phi-2","date":"2024-01-27T17:21:32.000Z","classes":["tokenizer"],"sentence":"phi-2: EOS declared as 50256 (was unset); generation that relied on the library default now stops on this token; BOS declared as 50256 (was unset) (2024-01-27).","keys":[{"file":"config","key":"bos_token_id","before":null,"after":50256},{"file":"config","key":"eos_token_id","before":null,"after":50256}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":{"before":null,"after":50256},"bos":{"before":null,"after":50256}},"fromSha":"313453cae69791d7650cda8f0134d4233f6f253c","toSha":"3f879ff35d20d910f0968d0b9b35a9fc074ecb27","commit":"https://huggingface.co/microsoft/phi-2/commit/3f879ff35d20d910f0968d0b9b35a9fc074ecb27","source":"backfill"},{"id":"microsoft/phi-2","date":"2024-01-27T17:21:23.000Z","classes":["tokenizer"],"sentence":"phi-2: EOS declared as 50256 (was unset); generation that relied on the library default now stops on this token; BOS declared as 50256 (was unset) (2024-01-27).","keys":[{"file":"generation","key":"bos_token_id","before":null,"after":50256},{"file":"generation","key":"eos_token_id","before":null,"after":50256}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":{"before":null,"after":50256},"bos":{"before":null,"after":50256}},"fromSha":"f05bc51ad3b1a729d5818514ddf091186398ac8c","toSha":"313453cae69791d7650cda8f0134d4233f6f253c","commit":"https://huggingface.co/microsoft/phi-2/commit/313453cae69791d7650cda8f0134d4233f6f253c","source":"backfill"},{"id":"jinaai/jina-embeddings-v2-base-code","date":"2024-01-25T05:13:22.000Z","classes":["structural"],"sentence":"jina-embeddings-v2-base-code: vocab_size changed from 61056 to 30528; the embedding and output head change shape, so adapters and merges made against the old vocabulary no longer load cleanly; derived parameters 138.2M to 114.8M (-23.4M, -0.05 GB of bf16 weights); declared dtype \"float16\" to \"float32\" (2024-01-25).","keys":[{"file":"config","key":"dtype","before":"float16","after":"float32"},{"file":"config","key":"vocab_size","before":61056,"after":30528}],"consequence":{"params":{"before":138218496,"after":114772992},"weightsGbBf16":{"before":0.276,"after":0.23},"trainPeakGb":{"before":23.492,"after":23.042},"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"c157a1f32981e53c466a5a7e27ef09b1cbdedee4","toSha":"27936d831769514f3107ecd67f9f202d74df59bf","commit":"https://huggingface.co/jinaai/jina-embeddings-v2-base-code/commit/27936d831769514f3107ecd67f9f202d74df59bf","source":"backfill"},{"id":"facebook/w2v-bert-2.0","date":"2024-01-16T09:35:02.000Z","classes":["structural"],"sentence":"w2v-bert-2.0: architectures [\"Wav2Vec2BERTForPreTraining\"] to [\"Wav2Vec2BertModel\"]; a different class is instantiated from the same id (2024-01-16).","keys":[{"file":"config","key":"architectures","before":["Wav2Vec2BERTForPreTraining"],"after":["Wav2Vec2BertModel"]}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"8353d2770d3c0d61f81b979d184a6caaa1404c63","toSha":"6b1c0b3a98343376f0cffd7cfbbe3ce4db45d459","commit":"https://huggingface.co/facebook/w2v-bert-2.0/commit/6b1c0b3a98343376f0cffd7cfbbe3ce4db45d459","source":"backfill"},{"id":"microsoft/phi-2","date":"2024-01-11T17:03:31.000Z","classes":["structural"],"sentence":"phi-2: MLP width unset to 10240, KV heads unset to 32; derived parameters 1.61B to 2.78B (+1.17B, +2.34 GB of bf16 weights); rope_theta unset to 10000; every position embedding changes (2024-01-11).","keys":[{"file":"config","key":"intermediate_size","before":null,"after":10240},{"file":"config","key":"model_type","before":"phi-msft","after":"phi"},{"file":"config","key":"num_key_value_heads","before":null,"after":32},{"file":"config","key":"rope_theta","before":null,"after":10000}],"consequence":{"params":{"before":1610456064,"after":2779520000},"weightsGbBf16":{"before":3.221,"after":5.559},"trainPeakGb":{"before":66.018,"after":274.321},"fingerprint":{"before":"d68ff415","after":"7fefd770"},"context":null,"eos":null,"bos":null},"fromSha":"d3186761bf5c4409f7679359284066c25ab668ee","toSha":"cb2f4533604d8b67de604e7df03bfe6f3ca22869","commit":"https://huggingface.co/microsoft/phi-2/commit/cb2f4533604d8b67de604e7df03bfe6f3ca22869","source":"backfill"},{"id":"mistralai/Mistral-7B-Instruct-v0.2","date":"2023-12-15T11:11:30.000Z","classes":["structural"],"sentence":"Mistral-7B-Instruct-v0.2: sliding_window 32768 to unset (2023-12-15).","keys":[{"file":"config","key":"sliding_window","before":32768,"after":null}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"cb35ee2c20437e8977f8b6b8ad6a7a636d520e88","toSha":"b70aa86578567ba3301b21c8a27bea4e8f6d6d61","commit":"https://huggingface.co/mistralai/Mistral-7B-Instruct-v0.2/commit/b70aa86578567ba3301b21c8a27bea4e8f6d6d61","source":"backfill"},{"id":"mistralai/Mistral-7B-Instruct-v0.2","date":"2023-12-14T22:36:52.000Z","classes":["structural"],"sentence":"Mistral-7B-Instruct-v0.2: sliding_window unset to 32768 (2023-12-14).","keys":[{"file":"config","key":"sliding_window","before":null,"after":32768}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"dca6e4b60aca009ed25ffa70c9bb65e46960a573","toSha":"cb35ee2c20437e8977f8b6b8ad6a7a636d520e88","commit":"https://huggingface.co/mistralai/Mistral-7B-Instruct-v0.2/commit/cb35ee2c20437e8977f8b6b8ad6a7a636d520e88","source":"backfill"},{"id":"microsoft/phi-2","date":"2023-12-13T23:20:53.000Z","classes":["structural"],"sentence":"phi-2: model_type \"phi\" to \"phi-msft\"; a different class is instantiated from the same id (2023-12-13).","keys":[{"file":"config","key":"model_type","before":"phi","after":"phi-msft"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"96f15add7380fc81e6ea20ab886edb5e917824f4","toSha":"40b720df64dcaa7a43b7c7a80715f077dfbd0d26","commit":"https://huggingface.co/microsoft/phi-2/commit/40b720df64dcaa7a43b7c7a80715f077dfbd0d26","source":"backfill"},{"id":"Qwen/Qwen-72B","date":"2023-11-30T07:04:19.000Z","classes":["tokenizer"],"sentence":"Qwen-72B: tokenizer model_max_length 8192 to 32768; truncation defaults change (2023-11-30).","keys":[{"file":"tokenizer","key":"model_max_length","before":8192,"after":32768}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"443804a5f97821a5d55810193459882df0c70759","toSha":"4c539df14494b0423c6644c10ea432aa7691ded8","commit":"https://huggingface.co/Qwen/Qwen-72B/commit/4c539df14494b0423c6644c10ea432aa7691ded8","source":"backfill"},{"id":"Qwen/Qwen-72B","date":"2023-11-28T16:03:12.000Z","classes":["structural"],"sentence":"Qwen-72B: context window 8192 to 32768 tokens (2023-11-28).","keys":[{"file":"config","key":"max_position_embeddings","before":8192,"after":32768}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":{"before":8192,"after":32768},"eos":null,"bos":null},"fromSha":"b4a48ab187995e0378b6e73e97d65bb166d2d807","toSha":"443804a5f97821a5d55810193459882df0c70759","commit":"https://huggingface.co/Qwen/Qwen-72B/commit/443804a5f97821a5d55810193459882df0c70759","source":"backfill"},{"id":"Qwen/Qwen-72B","date":"2023-11-27T15:01:07.000Z","classes":["structural"],"sentence":"Qwen-72B: declared dtype \"bfloat16\" to unset (2023-11-27).","keys":[{"file":"config","key":"dtype","before":"bfloat16","after":null}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"15a041f841af3b46ff0c1a31e2018c9dd2b38706","toSha":"b4a48ab187995e0378b6e73e97d65bb166d2d807","commit":"https://huggingface.co/Qwen/Qwen-72B/commit/b4a48ab187995e0378b6e73e97d65bb166d2d807","source":"backfill"},{"id":"openai/whisper-large-v3","date":"2023-11-16T11:44:34.000Z","classes":["tokenizer"],"sentence":"whisper-large-v3: EOS changed from 50256 to 50257; anything fine-tuned or served against the old value stops at a different token; BOS changed from 50256 to 50257; prompts built by the old tokenizer start with a different token (2023-11-16).","keys":[{"file":"config","key":"bos_token_id","before":50256,"after":50257},{"file":"config","key":"eos_token_id","before":50256,"after":50257}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":{"before":50256,"after":50257},"bos":{"before":50256,"after":50257}},"fromSha":"ca3e273640c81277b244489dc4a55c3eb9764436","toSha":"9ef08b92d5117306d28a6bcba2ca4ba407155f65","commit":"https://huggingface.co/openai/whisper-large-v3/commit/9ef08b92d5117306d28a6bcba2ca4ba407155f65","source":"backfill"},{"id":"openai/whisper-large-v3","date":"2023-11-07T18:53:52.000Z","classes":["tokenizer"],"sentence":"whisper-large-v3: pad token unset to \"<|endoftext|>\"; batched training masks a different id (2023-11-07).","keys":[{"file":"tokenizer","key":"pad_token","before":null,"after":"<|endoftext|>"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"8a2a8a3419314d1d44b89aec06271ca4394778f7","toSha":"d8411bd4e55c0bca39e60653a0fe26ae8591859a","commit":"https://huggingface.co/openai/whisper-large-v3/commit/d8411bd4e55c0bca39e60653a0fe26ae8591859a","source":"backfill"},{"id":"BAAI/bge-small-en-v1.5","date":"2023-11-02T10:47:51.000Z","classes":["tokenizer"],"sentence":"bge-small-en-v1.5: tokenizer model_max_length 1e+30 to 512; truncation defaults change (2023-11-02).","keys":[{"file":"tokenizer","key":"model_max_length","before":1e+30,"after":512}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"9b09f79fe618f3969e13e52f12c01de64145ebc0","toSha":"982532469af0dff5df8e70b38075b0940e863662","commit":"https://huggingface.co/BAAI/bge-small-en-v1.5/commit/982532469af0dff5df8e70b38075b0940e863662","source":"backfill"},{"id":"jinaai/jina-embeddings-v2-base-en","date":"2023-10-26T13:11:50.000Z","classes":["tokenizer"],"sentence":"jina-embeddings-v2-base-en: tokenizer model_max_length \"2147483648\" to 2147483648; truncation defaults change (2023-10-26).","keys":[{"file":"tokenizer","key":"model_max_length","before":"2147483648","after":2147483648}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"b59348ca8507000649a948b8260b2504482ccc9a","toSha":"0831c5fe3183a03fb20e24b8226fc4a0c7db23f6","commit":"https://huggingface.co/jinaai/jina-embeddings-v2-base-en/commit/0831c5fe3183a03fb20e24b8226fc4a0c7db23f6","source":"backfill"},{"id":"jinaai/jina-embeddings-v2-small-en","date":"2023-10-26T11:29:31.000Z","classes":["tokenizer"],"sentence":"jina-embeddings-v2-small-en: tokenizer model_max_length 8192 to 2147483648; truncation defaults change (2023-10-26).","keys":[{"file":"tokenizer","key":"model_max_length","before":8192,"after":2147483648}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"b63f8d36e31253b9b09c1a74b2e79efc55aaf47a","toSha":"1c993a952ef47cdd9e3576c1f22f935e5252f40c","commit":"https://huggingface.co/jinaai/jina-embeddings-v2-small-en/commit/1c993a952ef47cdd9e3576c1f22f935e5252f40c","source":"backfill"},{"id":"jinaai/jina-embeddings-v2-base-en","date":"2023-10-26T11:28:59.000Z","classes":["tokenizer"],"sentence":"jina-embeddings-v2-base-en: tokenizer model_max_length 8192 to \"2147483648\"; truncation defaults change (2023-10-26).","keys":[{"file":"tokenizer","key":"model_max_length","before":8192,"after":"2147483648"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"4409f4329786b15a551e3b855121ebf3ce76545a","toSha":"b59348ca8507000649a948b8260b2504482ccc9a","commit":"https://huggingface.co/jinaai/jina-embeddings-v2-base-en/commit/b59348ca8507000649a948b8260b2504482ccc9a","source":"backfill"},{"id":"jinaai/jina-embeddings-v2-small-en","date":"2023-10-19T20:43:20.000Z","classes":["structural"],"sentence":"jina-embeddings-v2-small-en: vocab_size changed from 30522 to 30528; the embedding and output head change shape, so adapters and merges made against the old vocabulary no longer load cleanly; derived parameters 32.4M to 32.4M (+3.1K, +0.00 GB of bf16 weights) (2023-10-19).","keys":[{"file":"config","key":"vocab_size","before":30522,"after":30528}],"consequence":{"params":{"before":32427008,"after":32430080},"weightsGbBf16":{"before":0.065,"after":0.065},"trainPeakGb":{"before":5.425,"after":5.425},"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"cb8edc5e5c8037c9c9613deb2d91ddc21c1d1b56","toSha":"8eac7ec07775c72f94b7012149aa5c87d80a73cf","commit":"https://huggingface.co/jinaai/jina-embeddings-v2-small-en/commit/8eac7ec07775c72f94b7012149aa5c87d80a73cf","source":"backfill"},{"id":"jinaai/jina-embeddings-v2-small-en","date":"2023-10-19T08:54:43.000Z","classes":["structural"],"sentence":"jina-embeddings-v2-small-en: vocab_size changed from 30528 to 30522; the embedding and output head change shape, so adapters and merges made against the old vocabulary no longer load cleanly; derived parameters 32.4M to 32.4M (-3.1K, -0.00 GB of bf16 weights) (2023-10-19).","keys":[{"file":"config","key":"vocab_size","before":30528,"after":30522}],"consequence":{"params":{"before":32430080,"after":32427008},"weightsGbBf16":{"before":0.065,"after":0.065},"trainPeakGb":{"before":5.425,"after":5.425},"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"2a070b375b8c211d7ddc60f5e65174583a0c2e32","toSha":"cb8edc5e5c8037c9c9613deb2d91ddc21c1d1b56","commit":"https://huggingface.co/jinaai/jina-embeddings-v2-small-en/commit/cb8edc5e5c8037c9c9613deb2d91ddc21c1d1b56","source":"backfill"},{"id":"jinaai/jina-embeddings-v2-small-en","date":"2023-10-04T13:56:33.000Z","classes":["structural"],"sentence":"jina-embeddings-v2-small-en: architectures [\"JBertForMaskedLM\"] to [\"JinaBertForMaskedLM\"]; a different class is instantiated from the same id (2023-10-04).","keys":[{"file":"config","key":"architectures","before":["JBertForMaskedLM"],"after":["JinaBertForMaskedLM"]}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"e19c7cb8d583f7e87bcddadfa49625c3a47f28c5","toSha":"f2416e24695db10826d50769c5dbfc421bde56d8","commit":"https://huggingface.co/jinaai/jina-embeddings-v2-small-en/commit/f2416e24695db10826d50769c5dbfc421bde56d8","source":"backfill"},{"id":"jinaai/jina-embeddings-v2-small-en","date":"2023-10-02T12:52:34.000Z","classes":["structural"],"sentence":"jina-embeddings-v2-small-en: derived parameters 28.5M to 32.4M (+3.9M, +0.01 GB of bf16 weights); context window 512 to 8192 tokens (2023-10-02).","keys":[{"file":"config","key":"max_position_embeddings","before":512,"after":8192}],"consequence":{"params":{"before":28497920,"after":32430080},"weightsGbBf16":{"before":0.057,"after":0.065},"trainPeakGb":{"before":1.748,"after":5.425},"fingerprint":null,"context":{"before":512,"after":8192},"eos":null,"bos":null},"fromSha":"e462ab66a9e43f837f2a117ae87f11a5b87f9a46","toSha":"e19c7cb8d583f7e87bcddadfa49625c3a47f28c5","commit":"https://huggingface.co/jinaai/jina-embeddings-v2-small-en/commit/e19c7cb8d583f7e87bcddadfa49625c3a47f28c5","source":"backfill"},{"id":"tiiuae/falcon-7b","date":"2023-09-29T14:32:19.000Z","classes":["structural","tokenizer"],"sentence":"falcon-7b: EOS changed from 2 to 11; anything fine-tuned or served against the old value stops at a different token; BOS changed from 1 to 11; prompts built by the old tokenizer start with a different token; architectures [\"RWForCausalLM\"] to [\"FalconForCausalLM\"]; a different class is instantiated from the same id (2023-09-29).","keys":[{"file":"config","key":"architectures","before":["RWForCausalLM"],"after":["FalconForCausalLM"]},{"file":"config","key":"model_type","before":"RefinedWebModel","after":"falcon"},{"file":"generation","key":"bos_token_id","before":1,"after":11},{"file":"generation","key":"eos_token_id","before":2,"after":11}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":{"before":2,"after":11},"bos":{"before":1,"after":11}},"fromSha":"378337427557d1df3e742264a2901a49f25d4eb1","toSha":"898df1396f35e447d5fe44e0a3ccaaaa69f30d36","commit":"https://huggingface.co/tiiuae/falcon-7b/commit/898df1396f35e447d5fe44e0a3ccaaaa69f30d36","source":"backfill"},{"id":"mistralai/Mistral-7B-v0.1","date":"2023-09-27T14:16:15.000Z","classes":["structural"],"sentence":"Mistral-7B-v0.1: context window 131072 to 32768 tokens (2023-09-27).","keys":[{"file":"config","key":"max_position_embeddings","before":131072,"after":32768}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":{"before":131072,"after":32768},"eos":null,"bos":null},"fromSha":"0e5a0e2d94a532654358ace70e2002cb93bbf4fb","toSha":"e836d8f71b5812f9fee65618453dc537c66bd82a","commit":"https://huggingface.co/mistralai/Mistral-7B-v0.1/commit/e836d8f71b5812f9fee65618453dc537c66bd82a","source":"backfill"},{"id":"mistralai/Mistral-7B-v0.1","date":"2023-09-26T08:18:26.000Z","classes":["tokenizer"],"sentence":"Mistral-7B-v0.1: tokenizer_class \"MistralTokenizer\" to \"LlamaTokenizer\" (2023-09-26).","keys":[{"file":"tokenizer","key":"tokenizer_class","before":"MistralTokenizer","after":"LlamaTokenizer"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"a8a923057860b4ab5000f2c83d3c4622be458aa0","toSha":"27f7f8c43bcdf706ef3fc74bf53a5897ce720eb2","commit":"https://huggingface.co/mistralai/Mistral-7B-v0.1/commit/27f7f8c43bcdf706ef3fc74bf53a5897ce720eb2","source":"backfill"},{"id":"mistralai/Mistral-7B-v0.1","date":"2023-09-25T07:52:54.000Z","classes":["structural","tokenizer"],"sentence":"Mistral-7B-v0.1: architectures [\"RafaleForCausalLM\"] to [\"MistralForCausalLM\"]; a different class is instantiated from the same id; tokenizer_class \"RafaleTokenizer\" to \"MistralTokenizer\" (2023-09-25).","keys":[{"file":"config","key":"architectures","before":["RafaleForCausalLM"],"after":["MistralForCausalLM"]},{"file":"config","key":"model_type","before":"rafale","after":"mistral"},{"file":"tokenizer","key":"tokenizer_class","before":"RafaleTokenizer","after":"MistralTokenizer"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"c2a147dc1311256b4072885a9ea67e4bf51bd926","toSha":"a8a923057860b4ab5000f2c83d3c4622be458aa0","commit":"https://huggingface.co/mistralai/Mistral-7B-v0.1/commit/a8a923057860b4ab5000f2c83d3c4622be458aa0","source":"backfill"},{"id":"BAAI/bge-base-en-v1.5","date":"2023-09-12T12:32:36.000Z","classes":["structural","tokenizer"],"sentence":"bge-base-en-v1.5: EOS 2 removed; generation falls back to whatever the library or another file declares; BOS 0 removed; vocab_size changed from 21128 to 30522; the embedding and output head change shape, so adapters and merges made against the old vocabulary no longer load cleanly (2023-09-12).","keys":[{"file":"config","key":"bos_token_id","before":0,"after":null},{"file":"config","key":"eos_token_id","before":2,"after":null},{"file":"config","key":"vocab_size","before":21128,"after":30522}],"consequence":{"params":{"before":101655552,"after":108870144},"weightsGbBf16":{"before":0.203,"after":0.218},"trainPeakGb":{"before":7.161,"after":7.3},"fingerprint":null,"context":null,"eos":{"before":2,"after":null},"bos":{"before":0,"after":null}},"fromSha":"6221169f04cebe20db3cbb3ca8d90fbd712770b6","toSha":"08e3d2c28e3886dfbcd98caebb447d9e23614bc0","commit":"https://huggingface.co/BAAI/bge-base-en-v1.5/commit/08e3d2c28e3886dfbcd98caebb447d9e23614bc0","source":"backfill"},{"id":"codellama/CodeLlama-7b-hf","date":"2023-08-28T17:28:57.000Z","classes":["structural"],"sentence":"CodeLlama-7b-hf: declared dtype \"float32\" to \"bfloat16\" (2023-08-28).","keys":[{"file":"config","key":"dtype","before":"float32","after":"bfloat16"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"0b6e5ef75b130071932322e0f3ca7b771c9f89fb","toSha":"ab8c8024db1b0b80226f31c32bce67b8d08fb87a","commit":"https://huggingface.co/codellama/CodeLlama-7b-hf/commit/ab8c8024db1b0b80226f31c32bce67b8d08fb87a","source":"backfill"},{"id":"codellama/CodeLlama-7b-hf","date":"2023-08-28T17:02:25.000Z","classes":["tokenizer"],"sentence":"CodeLlama-7b-hf: tokenizer_class \"LlamaTokenizer\" to \"CodeLlamaTokenizer\" (2023-08-28).","keys":[{"file":"tokenizer","key":"tokenizer_class","before":"LlamaTokenizer","after":"CodeLlamaTokenizer"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"be52f4ad322f5a47da121c761aeb5ba20ed77b17","toSha":"0b6e5ef75b130071932322e0f3ca7b771c9f89fb","commit":"https://huggingface.co/codellama/CodeLlama-7b-hf/commit/0b6e5ef75b130071932322e0f3ca7b771c9f89fb","source":"backfill"},{"id":"codellama/CodeLlama-7b-hf","date":"2023-08-25T09:59:03.000Z","classes":["structural"],"sentence":"CodeLlama-7b-hf: declared dtype \"bfloat16\" to \"float32\" (2023-08-25).","keys":[{"file":"config","key":"dtype","before":"bfloat16","after":"float32"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"ba606a6ffbf262e6e4217c2046f496748ec49eff","toSha":"be52f4ad322f5a47da121c761aeb5ba20ed77b17","commit":"https://huggingface.co/codellama/CodeLlama-7b-hf/commit/be52f4ad322f5a47da121c761aeb5ba20ed77b17","source":"backfill"},{"id":"tiiuae/falcon-7b","date":"2023-07-13T13:52:17.000Z","classes":["structural","tokenizer"],"sentence":"falcon-7b: EOS changed from 11 to 2; anything fine-tuned or served against the old value stops at a different token; BOS changed from 11 to 1; prompts built by the old tokenizer start with a different token; architectures [\"FalconForCausalLM\"] to [\"RWForCausalLM\"]; a different class is instantiated from the same id (2023-07-13).","keys":[{"file":"config","key":"architectures","before":["FalconForCausalLM"],"after":["RWForCausalLM"]},{"file":"config","key":"model_type","before":"falcon","after":"RefinedWebModel"},{"file":"generation","key":"bos_token_id","before":11,"after":1},{"file":"generation","key":"eos_token_id","before":11,"after":2}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":{"before":11,"after":2},"bos":{"before":11,"after":1}},"fromSha":"4e2d06f0a7c6370ebabbc30c6f59377ae8f73d76","toSha":"378337427557d1df3e742264a2901a49f25d4eb1","commit":"https://huggingface.co/tiiuae/falcon-7b/commit/378337427557d1df3e742264a2901a49f25d4eb1","source":"backfill"},{"id":"tiiuae/falcon-7b","date":"2023-07-12T21:33:43.000Z","classes":["structural","tokenizer"],"sentence":"falcon-7b: EOS changed from 2 to 11; anything fine-tuned or served against the old value stops at a different token; BOS changed from 1 to 11; prompts built by the old tokenizer start with a different token; architectures [\"RWForCausalLM\"] to [\"FalconForCausalLM\"]; a different class is instantiated from the same id (2023-07-12).","keys":[{"file":"config","key":"architectures","before":["RWForCausalLM"],"after":["FalconForCausalLM"]},{"file":"config","key":"model_type","before":"RefinedWebModel","after":"falcon"},{"file":"generation","key":"bos_token_id","before":1,"after":11},{"file":"generation","key":"eos_token_id","before":2,"after":11}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":{"before":2,"after":11},"bos":{"before":1,"after":11}},"fromSha":"2e3586947baae789afb6b8966bff3df919084bfb","toSha":"4e2d06f0a7c6370ebabbc30c6f59377ae8f73d76","commit":"https://huggingface.co/tiiuae/falcon-7b/commit/4e2d06f0a7c6370ebabbc30c6f59377ae8f73d76","source":"backfill"},{"id":"intfloat/multilingual-e5-small","date":"2023-07-12T09:14:02.000Z","classes":["tokenizer"],"sentence":"multilingual-e5-small: tokenizer model_max_length 1e+30 to 512; truncation defaults change (2023-07-12).","keys":[{"file":"tokenizer","key":"model_max_length","before":1e+30,"after":512}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"d2648e288f5fe1641aeab663a7fa6d1f0d1daff2","toSha":"960310a7742a4e0a250dd720466bed84fee8caac","commit":"https://huggingface.co/intfloat/multilingual-e5-small/commit/960310a7742a4e0a250dd720466bed84fee8caac","source":"backfill"},{"id":"tiiuae/falcon-7b","date":"2023-05-30T09:36:23.000Z","classes":["tokenizer"],"sentence":"falcon-7b: EOS changed from 2 to 11; anything fine-tuned or served against the old value stops at a different token; BOS changed from 1 to 11; prompts built by the old tokenizer start with a different token (2023-05-30).","keys":[{"file":"config","key":"bos_token_id","before":1,"after":11},{"file":"config","key":"eos_token_id","before":2,"after":11}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":{"before":2,"after":11},"bos":{"before":1,"after":11}},"fromSha":"b65214c38653907c2ed4cd160ffa84dc73a19c12","toSha":"f3185ff8325c74335b26f3e836d9050d86119fac","commit":"https://huggingface.co/tiiuae/falcon-7b/commit/f3185ff8325c74335b26f3e836d9050d86119fac","source":"backfill"},{"id":"tiiuae/falcon-7b","date":"2023-04-24T17:34:03.000Z","classes":["structural"],"sentence":"falcon-7b: model_type \"RefinedWeb\" to \"RefinedWebModel\"; a different class is instantiated from the same id (2023-04-24).","keys":[{"file":"config","key":"model_type","before":"RefinedWeb","after":"RefinedWebModel"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"80b61a69c52975a02865c89eca31a64bda2ac522","toSha":"956202585a08b264062407035023bd287efdae3b","commit":"https://huggingface.co/tiiuae/falcon-7b/commit/956202585a08b264062407035023bd287efdae3b","source":"backfill"},{"id":"EleutherAI/pythia-160m","date":"2023-04-03T18:39:04.000Z","classes":["tokenizer"],"sentence":"pythia-160m: EOS declared as \"<|endoftext|>\" (was unset); generation that relied on the library default now stops on this token; BOS declared as \"<|endoftext|>\" (was unset); tokenizer_class \"PreTrainedTokenizerFast\" to \"GPTNeoXTokenizer\" (2023-04-03).","keys":[{"file":"tokenizer","key":"bos_token","before":null,"after":"<|endoftext|>"},{"file":"tokenizer","key":"eos_token","before":null,"after":"<|endoftext|>"},{"file":"tokenizer","key":"tokenizer_class","before":"PreTrainedTokenizerFast","after":"GPTNeoXTokenizer"},{"file":"tokenizer","key":"unk_token","before":null,"after":"<|endoftext|>"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":{"before":null,"after":"<|endoftext|>"},"bos":{"before":null,"after":"<|endoftext|>"}},"fromSha":"610be53109eb1fb20ef697d2c139c5d3b07ee2f1","toSha":"4560d92fe0f8aafcf0839d845eec2c4f088f2496","commit":"https://huggingface.co/EleutherAI/pythia-160m/commit/4560d92fe0f8aafcf0839d845eec2c4f088f2496","source":"backfill"},{"id":"EleutherAI/pythia-160m","date":"2023-03-22T16:41:14.000Z","classes":["structural"],"sentence":"pythia-160m: declared dtype \"float32\" to \"float16\" (2023-03-22).","keys":[{"file":"config","key":"dtype","before":"float32","after":"float16"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"cd86582a32fbac1d74b94e7d851e3b3de0f074f5","toSha":"610be53109eb1fb20ef697d2c139c5d3b07ee2f1","commit":"https://huggingface.co/EleutherAI/pythia-160m/commit/610be53109eb1fb20ef697d2c139c5d3b07ee2f1","source":"backfill"},{"id":"EleutherAI/pythia-160m","date":"2023-03-22T16:34:29.000Z","classes":["structural"],"sentence":"pythia-160m: declared dtype \"float16\" to \"float32\" (2023-03-22).","keys":[{"file":"config","key":"dtype","before":"float16","after":"float32"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"9cab9b4640395ec6b208a3a6895e5a389420a25b","toSha":"cd86582a32fbac1d74b94e7d851e3b3de0f074f5","commit":"https://huggingface.co/EleutherAI/pythia-160m/commit/cd86582a32fbac1d74b94e7d851e3b3de0f074f5","source":"backfill"},{"id":"EleutherAI/pythia-160m","date":"2023-03-22T16:29:38.000Z","classes":["tokenizer"],"sentence":"pythia-160m: EOS \"<|endoftext|>\" removed; generation falls back to whatever the library or another file declares; BOS \"<|endoftext|>\" removed; tokenizer_class \"GPTNeoXTokenizer\" to \"PreTrainedTokenizerFast\" (2023-03-22).","keys":[{"file":"tokenizer","key":"bos_token","before":"<|endoftext|>","after":null},{"file":"tokenizer","key":"eos_token","before":"<|endoftext|>","after":null},{"file":"tokenizer","key":"tokenizer_class","before":"GPTNeoXTokenizer","after":"PreTrainedTokenizerFast"},{"file":"tokenizer","key":"unk_token","before":"<|endoftext|>","after":null}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":{"before":"<|endoftext|>","after":null},"bos":{"before":"<|endoftext|>","after":null}},"fromSha":"84b79d44e30834896564e1400db67ce25cd7263c","toSha":"9cab9b4640395ec6b208a3a6895e5a389420a25b","commit":"https://huggingface.co/EleutherAI/pythia-160m/commit/9cab9b4640395ec6b208a3a6895e5a389420a25b","source":"backfill"},{"id":"rinna/japanese-gpt-neox-small","date":"2023-03-20T08:00:35.000Z","classes":["structural"],"sentence":"japanese-gpt-neox-small: MLP width unset to 3072; model_type \"gpt-neox\" to \"gpt_neox\"; a different class is instantiated from the same id (2023-03-20).","keys":[{"file":"config","key":"intermediate_size","before":null,"after":3072},{"file":"config","key":"model_type","before":"gpt-neox","after":"gpt_neox"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":{"before":12.91,"after":42.716},"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"f3cdc9a0868ac3d6dbd8f7431f719dc07c5114e9","toSha":"d93a6e4cfd29b7a4cccc068c24f4afab9c218c45","commit":"https://huggingface.co/rinna/japanese-gpt-neox-small/commit/d93a6e4cfd29b7a4cccc068c24f4afab9c218c45","source":"backfill"},{"id":"bigscience/bloomz-560m","date":"2023-02-10T16:34:16.000Z","classes":["structural"],"sentence":"bloomz-560m: architectures [\"BloomModelForCausalLM\"] to [\"BloomForCausalLM\"]; a different class is instantiated from the same id (2023-02-10).","keys":[{"file":"config","key":"architectures","before":["BloomModelForCausalLM"],"after":["BloomForCausalLM"]}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"842d9731621a82db9e3cc92ca5ad7a84f709dc64","toSha":"64a6f1765615d2c38a7fe4474b8bff5a6c0e14e6","commit":"https://huggingface.co/bigscience/bloomz-560m/commit/64a6f1765615d2c38a7fe4474b8bff5a6c0e14e6","source":"backfill"},{"id":"bigscience/bloom-560m","date":"2023-02-10T15:08:30.000Z","classes":["structural"],"sentence":"bloom-560m: architectures [\"BloomModelForCausalLM\"] to [\"BloomForCausalLM\"]; a different class is instantiated from the same id (2023-02-10).","keys":[{"file":"config","key":"architectures","before":["BloomModelForCausalLM"],"after":["BloomForCausalLM"]}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"957a865efc636acac810ede2e891b8fb6e23f8e8","toSha":"e985a63cdc139290c5f700ff1929f0b5942cced2","commit":"https://huggingface.co/bigscience/bloom-560m/commit/e985a63cdc139290c5f700ff1929f0b5942cced2","source":"backfill"},{"id":"bigscience/bloomz-560m","date":"2023-02-10T14:32:35.000Z","classes":["structural"],"sentence":"bloomz-560m: architectures [\"BloomModel\"] to [\"BloomModelForCausalLM\"]; a different class is instantiated from the same id (2023-02-10).","keys":[{"file":"config","key":"architectures","before":["BloomModel"],"after":["BloomModelForCausalLM"]}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"4e825755a8883947d49929a73146b5fb631f09ed","toSha":"842d9731621a82db9e3cc92ca5ad7a84f709dc64","commit":"https://huggingface.co/bigscience/bloomz-560m/commit/842d9731621a82db9e3cc92ca5ad7a84f709dc64","source":"backfill"},{"id":"bigscience/bloom-560m","date":"2023-02-10T14:32:23.000Z","classes":["structural"],"sentence":"bloom-560m: architectures [\"BloomModel\"] to [\"BloomModelForCausalLM\"]; a different class is instantiated from the same id (2023-02-10).","keys":[{"file":"config","key":"architectures","before":["BloomModel"],"after":["BloomModelForCausalLM"]}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"148e7a5b6d054ef9e732088250d90ac78670890a","toSha":"957a865efc636acac810ede2e891b8fb6e23f8e8","commit":"https://huggingface.co/bigscience/bloom-560m/commit/957a865efc636acac810ede2e891b8fb6e23f8e8","source":"backfill"},{"id":"laion/CLIP-ViT-B-32-laion2B-s34B-b79K","date":"2023-01-22T22:52:38.000Z","classes":["tokenizer"],"sentence":"CLIP-ViT-B-32-laion2B-s34B-b79K: tokenizer model_max_length unset to 77; truncation defaults change; tokenizer_class unset to \"CLIPTokenizer\" (2023-01-22).","keys":[{"file":"tokenizer","key":"model_max_length","before":null,"after":77},{"file":"tokenizer","key":"tokenizer_class","before":null,"after":"CLIPTokenizer"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"cc77892181284701d11262109d3b85b4fd870813","toSha":"7937954a98e0d2fe2baa2aa0195f02d87d4e8389","commit":"https://huggingface.co/laion/CLIP-ViT-B-32-laion2B-s34B-b79K/commit/7937954a98e0d2fe2baa2aa0195f02d87d4e8389","source":"backfill"},{"id":"laion/CLIP-ViT-L-14-laion2B-s32B-b82K","date":"2023-01-22T22:52:03.000Z","classes":["tokenizer"],"sentence":"CLIP-ViT-L-14-laion2B-s32B-b82K: tokenizer model_max_length unset to 77; truncation defaults change; tokenizer_class unset to \"CLIPTokenizer\" (2023-01-22).","keys":[{"file":"tokenizer","key":"model_max_length","before":null,"after":77},{"file":"tokenizer","key":"tokenizer_class","before":null,"after":"CLIPTokenizer"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"5f9c1d932fff62b65782cd719cae9832ea4162a4","toSha":"b7028f054e49aaf2ab6775096a7fa14631591628","commit":"https://huggingface.co/laion/CLIP-ViT-L-14-laion2B-s32B-b82K/commit/b7028f054e49aaf2ab6775096a7fa14631591628","source":"backfill"},{"id":"google-t5/t5-base","date":"2022-12-12T16:09:56.000Z","classes":["structural"],"sentence":"t5-base: architectures [\"T5WithLMHeadModel\"] to [\"T5ForConditionalGeneration\"]; a different class is instantiated from the same id (2022-12-12).","keys":[{"file":"config","key":"architectures","before":["T5WithLMHeadModel"],"after":["T5ForConditionalGeneration"]}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"c1852398fbcceb89cd42afc33599f1edf42f3daa","toSha":"a4986abc64c580966eb8dd0ed61ceafaeb0d1d45","commit":"https://huggingface.co/google-t5/t5-base/commit/a4986abc64c580966eb8dd0ed61ceafaeb0d1d45","source":"backfill"},{"id":"google-t5/t5-small","date":"2022-12-12T16:06:17.000Z","classes":["structural"],"sentence":"t5-small: config.json was removed; whatever defaults it set now come from the library (2022-12-12).","keys":[{"file":"config","key":"(file)","before":"present","after":null}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"3ac06c3177fe685e9fba04b1166dde3a65efe70a","toSha":"c2bea447bc4e723126e9c5f126a9449c698168e1","commit":"https://huggingface.co/google-t5/t5-small/commit/c2bea447bc4e723126e9c5f126a9449c698168e1","source":"backfill"},{"id":"facebook/esm2_t33_650M_UR50D","date":"2022-12-01T18:58:37.000Z","classes":["tokenizer"],"sentence":"esm2_t33_650M_UR50D: tokenizer model_max_length unset to 1e+30; truncation defaults change (2022-12-01).","keys":[{"file":"tokenizer","key":"model_max_length","before":null,"after":1e+30}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"218058a409fe1a3e95a01fd034790ef8a6622cb5","toSha":"6123bfaf29ccd6aa250ac8861acc109ad5e7fb5d","commit":"https://huggingface.co/facebook/esm2_t33_650M_UR50D/commit/6123bfaf29ccd6aa250ac8861acc109ad5e7fb5d","source":"backfill"},{"id":"bigscience/bloom-560m","date":"2022-09-27T16:00:03.000Z","classes":["structural"],"sentence":"bloom-560m: context window 2048 to unset tokens (2022-09-27).","keys":[{"file":"config","key":"max_position_embeddings","before":2048,"after":null}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":{"before":2048,"after":null},"eos":null,"bos":null},"fromSha":"3e8a2f1867282e17817de0f8282c0ee199f45796","toSha":"148e7a5b6d054ef9e732088250d90ac78670890a","commit":"https://huggingface.co/bigscience/bloom-560m/commit/148e7a5b6d054ef9e732088250d90ac78670890a","source":"backfill"},{"id":"openai/clip-vit-large-patch14","date":"2022-09-09T17:34:57.000Z","classes":["structural"],"sentence":"clip-vit-large-patch14: config.json was removed; whatever defaults it set now come from the library (2022-09-09).","keys":[{"file":"config","key":"(file)","before":"present","after":null}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"352f696fa031fb4ce90e53dec6ddab4e48348543","toSha":"eb96ff977a370947ab00087870f497de35ec49ee","commit":"https://huggingface.co/openai/clip-vit-large-patch14/commit/eb96ff977a370947ab00087870f497de35ec49ee","source":"backfill"},{"id":"EleutherAI/gpt-neox-20b","date":"2022-08-09T11:39:40.000Z","classes":["tokenizer"],"sentence":"gpt-neox-20b: EOS changed from 2 to 0; anything fine-tuned or served against the old value stops at a different token (2022-08-09).","keys":[{"file":"config","key":"eos_token_id","before":2,"after":0}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":{"before":2,"after":0},"bos":null},"fromSha":"364ae95407723fadd1d47b023c1efb92a4d891c3","toSha":"3523781c8df75f7741687a4284f6f70e1afa12f4","commit":"https://huggingface.co/EleutherAI/gpt-neox-20b/commit/3523781c8df75f7741687a4284f6f70e1afa12f4","source":"backfill"},{"id":"bigscience/bloom-560m","date":"2022-07-12T14:48:30.000Z","classes":["structural"],"sentence":"bloom-560m: declared dtype \"float16\" to unset (2022-07-12).","keys":[{"file":"config","key":"dtype","before":"float16","after":null}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"5ce56ba7e04abb12bd68151538d6ae4155538f58","toSha":"7fdf1238d3cf6f3ea9f383ac44de1abffddf40aa","commit":"https://huggingface.co/bigscience/bloom-560m/commit/7fdf1238d3cf6f3ea9f383ac44de1abffddf40aa","source":"backfill"},{"id":"bigscience/bloom-560m","date":"2022-07-08T13:40:04.000Z","classes":["metadata"],"sentence":"bloom-560m: architectures declared as [\"BloomModel\"] (was unset) (2022-07-08).","keys":[{"file":"config","key":"architectures","before":null,"after":["BloomModel"]}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"0628d1bbfc2b2600d2d671d9d1dc0cc66c8d1867","toSha":"5ce56ba7e04abb12bd68151538d6ae4155538f58","commit":"https://huggingface.co/bigscience/bloom-560m/commit/5ce56ba7e04abb12bd68151538d6ae4155538f58","source":"backfill"},{"id":"bigscience/bloom-560m","date":"2022-06-22T17:47:06.000Z","classes":["tokenizer"],"sentence":"bloom-560m: tokenizer_class \"PreTrainedTokenizerFast\" to \"BloomTokenizerFast\" (2022-06-22).","keys":[{"file":"tokenizer","key":"tokenizer_class","before":"PreTrainedTokenizerFast","after":"BloomTokenizerFast"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"e9703d8bfd6d10dbdaeb9f57fd7c56410f576664","toSha":"0628d1bbfc2b2600d2d671d9d1dc0cc66c8d1867","commit":"https://huggingface.co/bigscience/bloom-560m/commit/0628d1bbfc2b2600d2d671d9d1dc0cc66c8d1867","source":"backfill"},{"id":"facebook/opt-125m","date":"2022-06-21T17:20:07.000Z","classes":["structural"],"sentence":"opt-125m: declared dtype \"float32\" to \"float16\" (2022-06-21).","keys":[{"file":"config","key":"dtype","before":"float32","after":"float16"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"80cabbb6d87253d8f161b42cbbb054a6a6d1c87e","toSha":"507a3991d874042a92e7581eb6e7cc7074b0c77e","commit":"https://huggingface.co/facebook/opt-125m/commit/507a3991d874042a92e7581eb6e7cc7074b0c77e","source":"backfill"},{"id":"facebook/opt-125m","date":"2022-06-03T08:52:00.000Z","classes":["structural"],"sentence":"opt-125m: declared dtype \"float16\" to \"float32\" (2022-06-03).","keys":[{"file":"config","key":"dtype","before":"float16","after":"float32"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"da9342600c033a6a5fbf1f3e9a75839107ecd887","toSha":"6fb08834a373ecaf6dea58677388dc8d0f8c23fe","commit":"https://huggingface.co/facebook/opt-125m/commit/6fb08834a373ecaf6dea58677388dc8d0f8c23fe","source":"backfill"},{"id":"bigscience/bloom-560m","date":"2022-05-28T18:44:45.000Z","classes":["tokenizer"],"sentence":"bloom-560m: EOS changed from 50256 to 2; anything fine-tuned or served against the old value stops at a different token; BOS changed from 50256 to 1; prompts built by the old tokenizer start with a different token; pad token unset to 3; batched training masks a different id (2022-05-28).","keys":[{"file":"config","key":"bos_token_id","before":50256,"after":1},{"file":"config","key":"eos_token_id","before":50256,"after":2},{"file":"config","key":"pad_token_id","before":null,"after":3}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":{"before":50256,"after":2},"bos":{"before":50256,"after":1}},"fromSha":"5fc13c0bee25a296f39ae3a6d4c8dc8dcb5f4c15","toSha":"8ad61ef2f1bf707dcec23c141de73b71e8b04827","commit":"https://huggingface.co/bigscience/bloom-560m/commit/8ad61ef2f1bf707dcec23c141de73b71e8b04827","source":"backfill"},{"id":"facebook/opt-125m","date":"2022-05-26T15:44:08.000Z","classes":["tokenizer"],"sentence":"opt-125m: BOS changed from 0 to 2; prompts built by the old tokenizer start with a different token (2022-05-26).","keys":[{"file":"config","key":"bos_token_id","before":0,"after":2}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":{"before":0,"after":2}},"fromSha":"1c370c849af40849404742f4d5ce1c2a0f82add2","toSha":"da9342600c033a6a5fbf1f3e9a75839107ecd887","commit":"https://huggingface.co/facebook/opt-125m/commit/da9342600c033a6a5fbf1f3e9a75839107ecd887","source":"backfill"},{"id":"facebook/opt-125m","date":"2022-05-25T20:19:35.000Z","classes":["tokenizer"],"sentence":"opt-125m: tokenizer_class \"GPT2Tokenizer\" to unset (2022-05-25).","keys":[{"file":"tokenizer","key":"tokenizer_class","before":"GPT2Tokenizer","after":null}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"c1c6d4a09608c0af95efdb024227a50e230d65c3","toSha":"1c370c849af40849404742f4d5ce1c2a0f82add2","commit":"https://huggingface.co/facebook/opt-125m/commit/1c370c849af40849404742f4d5ce1c2a0f82add2","source":"backfill"},{"id":"facebook/opt-125m","date":"2022-05-18T14:04:11.000Z","classes":["structural"],"sentence":"opt-125m: architectures [\"OPTModel\"] to [\"OPTForCausalLM\"]; a different class is instantiated from the same id (2022-05-18).","keys":[{"file":"config","key":"architectures","before":["OPTModel"],"after":["OPTForCausalLM"]}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"fae54032a2fff2773ba4f20b5263271c3c5eefab","toSha":"c1c6d4a09608c0af95efdb024227a50e230d65c3","commit":"https://huggingface.co/facebook/opt-125m/commit/c1c6d4a09608c0af95efdb024227a50e230d65c3","source":"backfill"},{"id":"facebook/opt-125m","date":"2022-05-17T14:58:09.000Z","classes":["structural"],"sentence":"opt-125m: architectures [\"OPTForCausalLM\"] to [\"OPTModel\"]; a different class is instantiated from the same id (2022-05-17).","keys":[{"file":"config","key":"architectures","before":["OPTForCausalLM"],"after":["OPTModel"]}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"74b53f9dd2c95e3338047711b9b30daa8bcb9525","toSha":"fae54032a2fff2773ba4f20b5263271c3c5eefab","commit":"https://huggingface.co/facebook/opt-125m/commit/fae54032a2fff2773ba4f20b5263271c3c5eefab","source":"backfill"},{"id":"facebook/opt-125m","date":"2022-05-13T08:42:48.000Z","classes":["structural"],"sentence":"opt-125m: architectures [\"OPTModel\"] to [\"OPTForCausalLM\"]; a different class is instantiated from the same id (2022-05-13).","keys":[{"file":"config","key":"architectures","before":["OPTModel"],"after":["OPTForCausalLM"]}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"04e12f0f0ce9ae1291797ad943acd63cbc609b5b","toSha":"74b53f9dd2c95e3338047711b9b30daa8bcb9525","commit":"https://huggingface.co/facebook/opt-125m/commit/74b53f9dd2c95e3338047711b9b30daa8bcb9525","source":"backfill"},{"id":"facebook/opt-125m","date":"2022-05-12T08:56:38.000Z","classes":["tokenizer"],"sentence":"opt-125m: EOS changed from \"<|endoftext|>\" to \"</s>\"; anything fine-tuned or served against the old value stops at a different token; BOS changed from \"<|endoftext|>\" to \"</s>\"; prompts built by the old tokenizer start with a different token; pad token unset to \"<pad>\"; batched training masks a different id (2022-05-12).","keys":[{"file":"tokenizer","key":"add_bos_token","before":null,"after":true},{"file":"tokenizer","key":"bos_token","before":"<|endoftext|>","after":"</s>"},{"file":"tokenizer","key":"eos_token","before":"<|endoftext|>","after":"</s>"},{"file":"tokenizer","key":"pad_token","before":null,"after":"<pad>"},{"file":"tokenizer","key":"unk_token","before":"<|endoftext|>","after":"</s>"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":{"before":"<|endoftext|>","after":"</s>"},"bos":{"before":"<|endoftext|>","after":"</s>"}},"fromSha":"3bbf6e43107ecc8e2afa89dfde0fa47cdb795312","toSha":"c29996b834cdfe3797401e7ebe368d58b4f9af61","commit":"https://huggingface.co/facebook/opt-125m/commit/c29996b834cdfe3797401e7ebe368d58b4f9af61","source":"backfill"},{"id":"sentence-transformers/all-MiniLM-L6-v2","date":"2021-08-30T20:00:57.000Z","classes":["tokenizer"],"sentence":"all-MiniLM-L6-v2: tokenizer model_max_length unset to 512; truncation defaults change (2021-08-30).","keys":[{"file":"tokenizer","key":"model_max_length","before":null,"after":512}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"2cfdf1ac2b3fc894d8239133174385c0baed9a17","toSha":"a4ce9694439b000ffd9ef4fe8becf6f62621689e","commit":"https://huggingface.co/sentence-transformers/all-MiniLM-L6-v2/commit/a4ce9694439b000ffd9ef4fe8becf6f62621689e","source":"backfill"},{"id":"facebook/bart-large-mnli","date":"2021-06-14T07:34:28.000Z","classes":["metadata"],"sentence":"bart-large-mnli: architectures declared as [\"BartForSequenceClassification\"] (was unset) (2021-06-14).","keys":[{"file":"config","key":"architectures","before":null,"after":["BartForSequenceClassification"]}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"e96ffdb4471996d791bfcbeb2447ec86186c35dd","toSha":"df7df4d2107d3882675bfd5ad50db1aa8dd7d7be","commit":"https://huggingface.co/facebook/bart-large-mnli/commit/df7df4d2107d3882675bfd5ad50db1aa8dd7d7be","source":"backfill"},{"id":"deepset/roberta-base-squad2","date":"2020-11-11T13:53:33.000Z","classes":["tokenizer"],"sentence":"roberta-base-squad2: EOS \"</s>\" removed; generation falls back to whatever the library or another file declares; BOS \"<s>\" removed; pad token \"<pad>\" to unset; batched training masks a different id (2020-11-11).","keys":[{"file":"tokenizer","key":"bos_token","before":"<s>","after":null},{"file":"tokenizer","key":"eos_token","before":"</s>","after":null},{"file":"tokenizer","key":"model_max_length","before":null,"after":512},{"file":"tokenizer","key":"pad_token","before":"<pad>","after":null},{"file":"tokenizer","key":"unk_token","before":"<unk>","after":null}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":{"before":"</s>","after":null},"bos":{"before":"<s>","after":null}},"fromSha":"596edb02b58a8e188cb86c2f0698e9bf46c83746","toSha":"deedc3e42208524e0df3d9149d1f26aa6934f05f","commit":"https://huggingface.co/deepset/roberta-base-squad2/commit/deedc3e42208524e0df3d9149d1f26aa6934f05f","source":"backfill"},{"id":"FacebookAI/xlm-roberta-base","date":"2020-04-24T16:07:22.000Z","classes":["tokenizer","metadata"],"sentence":"xlm-roberta-base: EOS declared as 2 (was unset); generation that relied on the library default now stops on this token; BOS declared as 0 (was unset); model_type declared as \"xlm-roberta\" (was unset) (2020-04-24).","keys":[{"file":"config","key":"bos_token_id","before":null,"after":0},{"file":"config","key":"eos_token_id","before":null,"after":2},{"file":"config","key":"model_type","before":null,"after":"xlm-roberta"},{"file":"config","key":"pad_token_id","before":null,"after":1}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":{"before":null,"after":2},"bos":{"before":null,"after":0}},"fromSha":"e17a4a08c15b2f6ce22b1883615bbfc001e3d6e9","toSha":"2cfd5ad1f384a79c5dd5d64b612fe16d3fa95292","commit":"https://huggingface.co/FacebookAI/xlm-roberta-base/commit/2cfd5ad1f384a79c5dd5d64b612fe16d3fa95292","source":"backfill"},{"id":"google-t5/t5-small","date":"2020-04-24T16:06:20.000Z","classes":["tokenizer","metadata"],"sentence":"t5-small: EOS declared as 1 (was unset); generation that relied on the library default now stops on this token; model_type declared as \"t5\" (was unset); pad token unset to 0; batched training masks a different id (2020-04-24).","keys":[{"file":"config","key":"eos_token_id","before":null,"after":1},{"file":"config","key":"model_type","before":null,"after":"t5"},{"file":"config","key":"pad_token_id","before":null,"after":0}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":{"before":null,"after":1},"bos":null},"fromSha":"00f126a8df80c3d2658ce9f1df7e573b740e3b9d","toSha":"3ac06c3177fe685e9fba04b1166dde3a65efe70a","commit":"https://huggingface.co/google-t5/t5-small/commit/3ac06c3177fe685e9fba04b1166dde3a65efe70a","source":"backfill"},{"id":"google-t5/t5-base","date":"2020-04-24T16:06:15.000Z","classes":["tokenizer","metadata"],"sentence":"t5-base: EOS declared as 1 (was unset); generation that relied on the library default now stops on this token; model_type declared as \"t5\" (was unset); pad token unset to 0; batched training masks a different id (2020-04-24).","keys":[{"file":"config","key":"eos_token_id","before":null,"after":1},{"file":"config","key":"model_type","before":null,"after":"t5"},{"file":"config","key":"pad_token_id","before":null,"after":0}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":{"before":null,"after":1},"bos":null},"fromSha":"f611a04e050c6b3d21560384b0bf4d18479027ed","toSha":"c1852398fbcceb89cd42afc33599f1edf42f3daa","commit":"https://huggingface.co/google-t5/t5-base/commit/c1852398fbcceb89cd42afc33599f1edf42f3daa","source":"backfill"},{"id":"FacebookAI/roberta-large","date":"2020-04-24T16:05:42.000Z","classes":["tokenizer","metadata"],"sentence":"roberta-large: EOS declared as 2 (was unset); generation that relied on the library default now stops on this token; BOS declared as 0 (was unset); model_type declared as \"roberta\" (was unset) (2020-04-24).","keys":[{"file":"config","key":"bos_token_id","before":null,"after":0},{"file":"config","key":"eos_token_id","before":null,"after":2},{"file":"config","key":"model_type","before":null,"after":"roberta"},{"file":"config","key":"pad_token_id","before":null,"after":1}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":{"before":null,"after":2},"bos":{"before":null,"after":0}},"fromSha":"66311c774099612859a28344a022402a379463a9","toSha":"5674b5c411cae16c94622d617b434cbd5e390d05","commit":"https://huggingface.co/FacebookAI/roberta-large/commit/5674b5c411cae16c94622d617b434cbd5e390d05","source":"backfill"},{"id":"FacebookAI/roberta-base","date":"2020-04-24T16:05:38.000Z","classes":["tokenizer","metadata"],"sentence":"roberta-base: EOS declared as 2 (was unset); generation that relied on the library default now stops on this token; BOS declared as 0 (was unset); model_type declared as \"roberta\" (was unset) (2020-04-24).","keys":[{"file":"config","key":"bos_token_id","before":null,"after":0},{"file":"config","key":"eos_token_id","before":null,"after":2},{"file":"config","key":"model_type","before":null,"after":"roberta"},{"file":"config","key":"pad_token_id","before":null,"after":1}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":{"before":null,"after":2},"bos":{"before":null,"after":0}},"fromSha":"2ecd732fb792ee9de217a2f7a7154e044cf3d782","toSha":"47a52b98637fc8df2bf37a7b785cea13ad81ba94","commit":"https://huggingface.co/FacebookAI/roberta-base/commit/47a52b98637fc8df2bf37a7b785cea13ad81ba94","source":"backfill"},{"id":"openai-community/gpt2-large","date":"2020-04-24T16:00:20.000Z","classes":["tokenizer","metadata"],"sentence":"gpt2-large: EOS declared as 50256 (was unset); generation that relied on the library default now stops on this token; BOS declared as 50256 (was unset); model_type declared as \"gpt2\" (was unset) (2020-04-24).","keys":[{"file":"config","key":"bos_token_id","before":null,"after":50256},{"file":"config","key":"eos_token_id","before":null,"after":50256},{"file":"config","key":"model_type","before":null,"after":"gpt2"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":{"before":null,"after":50256},"bos":{"before":null,"after":50256}},"fromSha":"69349024845f2bb2207abd6e8098fe63583a703b","toSha":"5b4837031641fd4d080e59711d9beb89ce35d5d3","commit":"https://huggingface.co/openai-community/gpt2-large/commit/5b4837031641fd4d080e59711d9beb89ce35d5d3","source":"backfill"},{"id":"openai-community/gpt2","date":"2020-04-24T16:00:19.000Z","classes":["tokenizer","metadata"],"sentence":"gpt2: EOS declared as 50256 (was unset); generation that relied on the library default now stops on this token; BOS declared as 50256 (was unset); model_type declared as \"gpt2\" (was unset) (2020-04-24).","keys":[{"file":"config","key":"bos_token_id","before":null,"after":50256},{"file":"config","key":"eos_token_id","before":null,"after":50256},{"file":"config","key":"model_type","before":null,"after":"gpt2"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":{"before":null,"after":50256},"bos":{"before":null,"after":50256}},"fromSha":"8bfcaddd2b23a8f06826be2d8fec04be6674aee1","toSha":"26b7c7a64cce1b361f6083462bc27c3ca382dd44","commit":"https://huggingface.co/openai-community/gpt2/commit/26b7c7a64cce1b361f6083462bc27c3ca382dd44","source":"backfill"},{"id":"facebook/bart-large-mnli","date":"2020-04-24T15:58:44.000Z","classes":["tokenizer"],"sentence":"bart-large-mnli: EOS declared as 2 (was unset); generation that relied on the library default now stops on this token; BOS declared as 0 (was unset); pad token unset to 1; batched training masks a different id (2020-04-24).","keys":[{"file":"config","key":"bos_token_id","before":null,"after":0},{"file":"config","key":"eos_token_id","before":null,"after":2},{"file":"config","key":"pad_token_id","before":null,"after":1}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":{"before":null,"after":2},"bos":{"before":null,"after":0}},"fromSha":"a9aaa74e9eb0c21b338955c15bb56489b869e193","toSha":"048916aef0957008cf5628d087c4a1e794704f73","commit":"https://huggingface.co/facebook/bart-large-mnli/commit/048916aef0957008cf5628d087c4a1e794704f73","source":"backfill"},{"id":"distilbert/distilgpt2","date":"2020-04-24T15:58:08.000Z","classes":["tokenizer","metadata"],"sentence":"distilgpt2: EOS declared as 50256 (was unset); generation that relied on the library default now stops on this token; BOS declared as 50256 (was unset); model_type declared as \"gpt2\" (was unset) (2020-04-24).","keys":[{"file":"config","key":"bos_token_id","before":null,"after":50256},{"file":"config","key":"eos_token_id","before":null,"after":50256},{"file":"config","key":"model_type","before":null,"after":"gpt2"}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":{"before":null,"after":50256},"bos":{"before":null,"after":50256}},"fromSha":"51ab088a71e3e57d331248881854feb0865fed4e","toSha":"b09826e12b0da3e8b95026b6866164fd4dee6511","commit":"https://huggingface.co/distilbert/distilgpt2/commit/b09826e12b0da3e8b95026b6866164fd4dee6511","source":"backfill"},{"id":"distilbert/distilbert-base-uncased-finetuned-sst-2-english","date":"2020-04-24T15:58:05.000Z","classes":["tokenizer","metadata"],"sentence":"distilbert-base-uncased-finetuned-sst-2-english: model_type declared as \"distilbert\" (was unset); pad token unset to 0; batched training masks a different id (2020-04-24).","keys":[{"file":"config","key":"model_type","before":null,"after":"distilbert"},{"file":"config","key":"pad_token_id","before":null,"after":0}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"d9fd1c824e6530b38e19b0ebcf94d67253e5ddb0","toSha":"9e7b6fe74288a25833caa9591d1b2dc70acc628f","commit":"https://huggingface.co/distilbert/distilbert-base-uncased-finetuned-sst-2-english/commit/9e7b6fe74288a25833caa9591d1b2dc70acc628f","source":"backfill"},{"id":"distilbert/distilbert-base-uncased","date":"2020-04-24T15:58:01.000Z","classes":["tokenizer","metadata"],"sentence":"distilbert-base-uncased: model_type declared as \"distilbert\" (was unset); pad token unset to 0; batched training masks a different id (2020-04-24).","keys":[{"file":"config","key":"model_type","before":null,"after":"distilbert"},{"file":"config","key":"pad_token_id","before":null,"after":0}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"1b40ea272d7059083b6d8b90273924bcbdcdb4b5","toSha":"554bcd28aa7be76e65e51825031e918f5346964f","commit":"https://huggingface.co/distilbert/distilbert-base-uncased/commit/554bcd28aa7be76e65e51825031e918f5346964f","source":"backfill"},{"id":"deepset/roberta-base-squad2","date":"2020-04-24T15:57:45.000Z","classes":["tokenizer"],"sentence":"roberta-base-squad2: EOS declared as 2 (was unset); generation that relied on the library default now stops on this token; pad token unset to 1; batched training masks a different id (2020-04-24).","keys":[{"file":"config","key":"eos_token_id","before":null,"after":2},{"file":"config","key":"pad_token_id","before":null,"after":1}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":{"before":null,"after":2},"bos":null},"fromSha":"784fa08ee01e943e70962a88f9935514ccf3964e","toSha":"4aa045039343a5dbb88bab82d8e6a3529ca8a1bc","commit":"https://huggingface.co/deepset/roberta-base-squad2/commit/4aa045039343a5dbb88bab82d8e6a3529ca8a1bc","source":"backfill"},{"id":"dbmdz/bert-large-cased-finetuned-conll03-english","date":"2020-04-24T15:57:15.000Z","classes":["tokenizer","metadata"],"sentence":"bert-large-cased-finetuned-conll03-english: model_type declared as \"bert\" (was unset); pad token unset to 0; batched training masks a different id (2020-04-24).","keys":[{"file":"config","key":"model_type","before":null,"after":"bert"},{"file":"config","key":"pad_token_id","before":null,"after":0}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"c15d1bd2bc7ec48610e10ead861423a4f3717056","toSha":"7405d6fca06856aed5b135a9e6254853c36e6a45","commit":"https://huggingface.co/dbmdz/bert-large-cased-finetuned-conll03-english/commit/7405d6fca06856aed5b135a9e6254853c36e6a45","source":"backfill"},{"id":"almanach/camembert-base","date":"2020-04-24T15:55:46.000Z","classes":["tokenizer","metadata"],"sentence":"camembert-base: model_type declared as \"camembert\" (was unset); pad token unset to 1; batched training masks a different id (2020-04-24).","keys":[{"file":"config","key":"model_type","before":null,"after":"camembert"},{"file":"config","key":"pad_token_id","before":null,"after":1}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"6f36eb8a4d9bf54b2b3fc7977e2b3aed71b604fd","toSha":"81407e9a85a3505c39252318c8caa572e20f385d","commit":"https://huggingface.co/almanach/camembert-base/commit/81407e9a85a3505c39252318c8caa572e20f385d","source":"backfill"},{"id":"google-bert/bert-base-uncased","date":"2020-04-24T15:55:13.000Z","classes":["tokenizer","metadata"],"sentence":"bert-base-uncased: model_type declared as \"bert\" (was unset); pad token unset to 0; batched training masks a different id (2020-04-24).","keys":[{"file":"config","key":"model_type","before":null,"after":"bert"},{"file":"config","key":"pad_token_id","before":null,"after":0}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"941c1ed3859432b6399878f50a93a7afe8394b1a","toSha":"5b16e96ccd6a42974d9bb1761b6506b896b83537","commit":"https://huggingface.co/google-bert/bert-base-uncased/commit/5b16e96ccd6a42974d9bb1761b6506b896b83537","source":"backfill"},{"id":"google-bert/bert-base-multilingual-uncased","date":"2020-04-24T15:55:11.000Z","classes":["tokenizer","metadata"],"sentence":"bert-base-multilingual-uncased: model_type declared as \"bert\" (was unset); pad token unset to 0; batched training masks a different id (2020-04-24).","keys":[{"file":"config","key":"model_type","before":null,"after":"bert"},{"file":"config","key":"pad_token_id","before":null,"after":0}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"cf0d44a0ce3e24e426cbc6979819deed2b329c69","toSha":"e6a669dea4f3905a38bed042ae9fa07b80f8d880","commit":"https://huggingface.co/google-bert/bert-base-multilingual-uncased/commit/e6a669dea4f3905a38bed042ae9fa07b80f8d880","source":"backfill"},{"id":"google-bert/bert-base-cased","date":"2020-04-24T15:54:55.000Z","classes":["tokenizer","metadata"],"sentence":"bert-base-cased: model_type declared as \"bert\" (was unset); pad token unset to 0; batched training masks a different id (2020-04-24).","keys":[{"file":"config","key":"model_type","before":null,"after":"bert"},{"file":"config","key":"pad_token_id","before":null,"after":0}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"38e5f9c0af8586e2e980a13387ede80a4461d22e","toSha":"6c38be42d6b3ed125933ca95fd0165fb9df8e414","commit":"https://huggingface.co/google-bert/bert-base-cased/commit/6c38be42d6b3ed125933ca95fd0165fb9df8e414","source":"backfill"},{"id":"google-t5/t5-small","date":"2020-03-26T08:05:44.000Z","classes":["structural"],"sentence":"t5-small: config.json was removed; whatever defaults it set now come from the library (2020-03-26).","keys":[{"file":"config","key":"(file)","before":"present","after":null}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"e87f2a02b2fa563300e788b36ecdaab4fc6bf892","toSha":"8c5d40ac989ba52e99dbdd964f6f4e5381f080c7","commit":"https://huggingface.co/google-t5/t5-small/commit/8c5d40ac989ba52e99dbdd964f6f4e5381f080c7","source":"backfill"},{"id":"deepset/roberta-base-squad2","date":"2020-03-23T21:43:57.000Z","classes":["tokenizer"],"sentence":"roberta-base-squad2: pad token 0 to unset; batched training masks a different id (2020-03-23).","keys":[{"file":"config","key":"pad_token_id","before":0,"after":null}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"8f39c9f238c22cd65cc499d0f69224742c6f30b7","toSha":"784fa08ee01e943e70962a88f9935514ccf3964e","commit":"https://huggingface.co/deepset/roberta-base-squad2/commit/784fa08ee01e943e70962a88f9935514ccf3964e","source":"backfill"},{"id":"almanach/camembert-base","date":"2020-03-23T21:41:27.000Z","classes":["tokenizer"],"sentence":"camembert-base: EOS declared as 6 (was unset); generation that relied on the library default now stops on this token; BOS declared as 5 (was unset) (2020-03-23).","keys":[{"file":"config","key":"bos_token_id","before":null,"after":5},{"file":"config","key":"eos_token_id","before":null,"after":6}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":{"before":null,"after":6},"bos":{"before":null,"after":5}},"fromSha":"c518dfe4b3cb3f82803e51c284f26d50633f0fc0","toSha":"6f36eb8a4d9bf54b2b3fc7977e2b3aed71b604fd","commit":"https://huggingface.co/almanach/camembert-base/commit/6f36eb8a4d9bf54b2b3fc7977e2b3aed71b604fd","source":"backfill"},{"id":"facebook/bart-large-mnli","date":"2020-03-14T15:56:14.000Z","classes":["tokenizer"],"sentence":"bart-large-mnli: EOS 2 removed; generation falls back to whatever the library or another file declares; BOS 0 removed; pad token 1 to unset; batched training masks a different id (2020-03-14).","keys":[{"file":"config","key":"bos_token_id","before":0,"after":null},{"file":"config","key":"eos_token_id","before":2,"after":null},{"file":"config","key":"pad_token_id","before":1,"after":null}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":{"before":2,"after":null},"bos":{"before":0,"after":null}},"fromSha":"e086d2bf298037fb49e60ad3a9640953aed84fde","toSha":"a9aaa74e9eb0c21b338955c15bb56489b869e193","commit":"https://huggingface.co/facebook/bart-large-mnli/commit/a9aaa74e9eb0c21b338955c15bb56489b869e193","source":"backfill"},{"id":"FacebookAI/xlm-roberta-base","date":"2020-01-31T23:00:32.000Z","classes":["metadata"],"sentence":"xlm-roberta-base: architectures declared as [\"XLMRobertaForMaskedLM\"] (was unset) (2020-01-31).","keys":[{"file":"config","key":"architectures","before":null,"after":["XLMRobertaForMaskedLM"]}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"d1a321d00082e2ef487a7b6845dea4efedd7aba4","toSha":"e17a4a08c15b2f6ce22b1883615bbfc001e3d6e9","commit":"https://huggingface.co/FacebookAI/xlm-roberta-base/commit/e17a4a08c15b2f6ce22b1883615bbfc001e3d6e9","source":"backfill"},{"id":"google-t5/t5-base","date":"2020-01-31T23:00:30.000Z","classes":["metadata"],"sentence":"t5-base: architectures declared as [\"T5WithLMHeadModel\"] (was unset) (2020-01-31).","keys":[{"file":"config","key":"architectures","before":null,"after":["T5WithLMHeadModel"]}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"666f4ee1ab305b1835b95ee83c183c4d0948bf2e","toSha":"06e82b5732351916b539cacf91e0d644db49c864","commit":"https://huggingface.co/google-t5/t5-base/commit/06e82b5732351916b539cacf91e0d644db49c864","source":"backfill"},{"id":"google-t5/t5-small","date":"2020-01-31T23:00:30.000Z","classes":["metadata"],"sentence":"t5-small: architectures declared as [\"T5WithLMHeadModel\"] (was unset) (2020-01-31).","keys":[{"file":"config","key":"architectures","before":null,"after":["T5WithLMHeadModel"]}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"c9c2c8f7fe6aa9ce37f61418d82a01d25cfac393","toSha":"e87f2a02b2fa563300e788b36ecdaab4fc6bf892","commit":"https://huggingface.co/google-t5/t5-small/commit/e87f2a02b2fa563300e788b36ecdaab4fc6bf892","source":"backfill"},{"id":"FacebookAI/roberta-base","date":"2020-01-31T23:00:29.000Z","classes":["metadata"],"sentence":"roberta-base: architectures declared as [\"RobertaForMaskedLM\"] (was unset) (2020-01-31).","keys":[{"file":"config","key":"architectures","before":null,"after":["RobertaForMaskedLM"]}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"e306078ce88386f7afbfa6e79a76d19b7e87f0c8","toSha":"2ecd732fb792ee9de217a2f7a7154e044cf3d782","commit":"https://huggingface.co/FacebookAI/roberta-base/commit/2ecd732fb792ee9de217a2f7a7154e044cf3d782","source":"backfill"},{"id":"FacebookAI/roberta-large","date":"2020-01-31T23:00:29.000Z","classes":["metadata"],"sentence":"roberta-large: architectures declared as [\"RobertaForMaskedLM\"] (was unset) (2020-01-31).","keys":[{"file":"config","key":"architectures","before":null,"after":["RobertaForMaskedLM"]}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"7cca3b740828005857c8d56f2b99267396329e79","toSha":"66311c774099612859a28344a022402a379463a9","commit":"https://huggingface.co/FacebookAI/roberta-large/commit/66311c774099612859a28344a022402a379463a9","source":"backfill"},{"id":"openai-community/gpt2","date":"2020-01-31T23:00:25.000Z","classes":["metadata"],"sentence":"gpt2: architectures declared as [\"GPT2LMHeadModel\"] (was unset) (2020-01-31).","keys":[{"file":"config","key":"architectures","before":null,"after":["GPT2LMHeadModel"]}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"703af3447380a4fdf35eb5c2c13b281dc3fd2cad","toSha":"8bfcaddd2b23a8f06826be2d8fec04be6674aee1","commit":"https://huggingface.co/openai-community/gpt2/commit/8bfcaddd2b23a8f06826be2d8fec04be6674aee1","source":"backfill"},{"id":"openai-community/gpt2-large","date":"2020-01-31T23:00:25.000Z","classes":["metadata"],"sentence":"gpt2-large: architectures declared as [\"GPT2LMHeadModel\"] (was unset) (2020-01-31).","keys":[{"file":"config","key":"architectures","before":null,"after":["GPT2LMHeadModel"]}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"a1a649487a4fbb683b532ab19a826cca540aa0f7","toSha":"69349024845f2bb2207abd6e8098fe63583a703b","commit":"https://huggingface.co/openai-community/gpt2-large/commit/69349024845f2bb2207abd6e8098fe63583a703b","source":"backfill"},{"id":"deepset/roberta-base-squad2","date":"2020-01-31T23:00:24.000Z","classes":["metadata"],"sentence":"roberta-base-squad2: architectures declared as [\"RobertaForQuestionAnswering\"] (was unset) (2020-01-31).","keys":[{"file":"config","key":"architectures","before":null,"after":["RobertaForQuestionAnswering"]}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"0eb280af961391cdfbc39e6eaebc7878ba7b7841","toSha":"8f39c9f238c22cd65cc499d0f69224742c6f30b7","commit":"https://huggingface.co/deepset/roberta-base-squad2/commit/8f39c9f238c22cd65cc499d0f69224742c6f30b7","source":"backfill"},{"id":"distilbert/distilbert-base-uncased","date":"2020-01-31T23:00:24.000Z","classes":["metadata"],"sentence":"distilbert-base-uncased: architectures declared as [\"DistilBertForMaskedLM\"] (was unset) (2020-01-31).","keys":[{"file":"config","key":"architectures","before":null,"after":["DistilBertForMaskedLM"]}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"713fdbc4b397e2ca0bc96dbd6c552869ea5829cd","toSha":"1b40ea272d7059083b6d8b90273924bcbdcdb4b5","commit":"https://huggingface.co/distilbert/distilbert-base-uncased/commit/1b40ea272d7059083b6d8b90273924bcbdcdb4b5","source":"backfill"},{"id":"distilbert/distilbert-base-uncased-finetuned-sst-2-english","date":"2020-01-31T23:00:24.000Z","classes":["metadata"],"sentence":"distilbert-base-uncased-finetuned-sst-2-english: architectures declared as [\"DistilBertForSequenceClassification\"] (was unset) (2020-01-31).","keys":[{"file":"config","key":"architectures","before":null,"after":["DistilBertForSequenceClassification"]}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"32d9c76413a46e99c76b159141864c48dd2e12d8","toSha":"d9fd1c824e6530b38e19b0ebcf94d67253e5ddb0","commit":"https://huggingface.co/distilbert/distilbert-base-uncased-finetuned-sst-2-english/commit/d9fd1c824e6530b38e19b0ebcf94d67253e5ddb0","source":"backfill"},{"id":"distilbert/distilgpt2","date":"2020-01-31T23:00:24.000Z","classes":["metadata"],"sentence":"distilgpt2: architectures declared as [\"GPT2LMHeadModel\"] (was unset) (2020-01-31).","keys":[{"file":"config","key":"architectures","before":null,"after":["GPT2LMHeadModel"]}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"07999848cf39ac1751b8c9a9854a30e020f6ebb5","toSha":"51ab088a71e3e57d331248881854feb0865fed4e","commit":"https://huggingface.co/distilbert/distilgpt2/commit/51ab088a71e3e57d331248881854feb0865fed4e","source":"backfill"},{"id":"dbmdz/bert-large-cased-finetuned-conll03-english","date":"2020-01-31T23:00:21.000Z","classes":["metadata"],"sentence":"bert-large-cased-finetuned-conll03-english: architectures declared as [\"BertForTokenClassification\"] (was unset) (2020-01-31).","keys":[{"file":"config","key":"architectures","before":null,"after":["BertForTokenClassification"]}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"c842729cdaab14eb8dae053bbb75c03920d5b5bc","toSha":"e2141104b0d8201ad9b0b7dcad1c7da06356352b","commit":"https://huggingface.co/dbmdz/bert-large-cased-finetuned-conll03-english/commit/e2141104b0d8201ad9b0b7dcad1c7da06356352b","source":"backfill"},{"id":"almanach/camembert-base","date":"2020-01-31T23:00:18.000Z","classes":["metadata"],"sentence":"camembert-base: architectures declared as [\"CamembertForMaskedLM\"] (was unset) (2020-01-31).","keys":[{"file":"config","key":"architectures","before":null,"after":["CamembertForMaskedLM"]}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"d5643396ad524a48deab8f55f07a156f15f27d5d","toSha":"c518dfe4b3cb3f82803e51c284f26d50633f0fc0","commit":"https://huggingface.co/almanach/camembert-base/commit/c518dfe4b3cb3f82803e51c284f26d50633f0fc0","source":"backfill"},{"id":"google-bert/bert-base-multilingual-uncased","date":"2020-01-31T23:00:18.000Z","classes":["metadata"],"sentence":"bert-base-multilingual-uncased: architectures declared as [\"BertForMaskedLM\"] (was unset) (2020-01-31).","keys":[{"file":"config","key":"architectures","before":null,"after":["BertForMaskedLM"]}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"19108269e89bf33f830e99a527c7521965f80ec5","toSha":"cf0d44a0ce3e24e426cbc6979819deed2b329c69","commit":"https://huggingface.co/google-bert/bert-base-multilingual-uncased/commit/cf0d44a0ce3e24e426cbc6979819deed2b329c69","source":"backfill"},{"id":"google-bert/bert-base-uncased","date":"2020-01-31T23:00:18.000Z","classes":["metadata"],"sentence":"bert-base-uncased: architectures declared as [\"BertForMaskedLM\"] (was unset) (2020-01-31).","keys":[{"file":"config","key":"architectures","before":null,"after":["BertForMaskedLM"]}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"bb3c1c3256d2598217df9889a14a2e811587891d","toSha":"941c1ed3859432b6399878f50a93a7afe8394b1a","commit":"https://huggingface.co/google-bert/bert-base-uncased/commit/941c1ed3859432b6399878f50a93a7afe8394b1a","source":"backfill"},{"id":"google-bert/bert-base-cased","date":"2020-01-31T23:00:17.000Z","classes":["metadata"],"sentence":"bert-base-cased: architectures declared as [\"BertForMaskedLM\"] (was unset) (2020-01-31).","keys":[{"file":"config","key":"architectures","before":null,"after":["BertForMaskedLM"]}],"consequence":{"params":null,"weightsGbBf16":null,"trainPeakGb":null,"fingerprint":null,"context":null,"eos":null,"bos":null},"fromSha":"2d2cf4f954ac07fa8972025022a9369398df9356","toSha":"38e5f9c0af8586e2e980a13387ede80a4461d22e","commit":"https://huggingface.co/google-bert/bert-base-cased/commit/38e5f9c0af8586e2e980a13387ede80a4461d22e","source":"backfill"},{"id":"FacebookAI/roberta-base","date":"2019-09-26T17:03:17.000Z","classes":["structural"],"sentence":"roberta-base: derived parameters 124.0M to 124.0M (+1.5K, +0.00 GB of bf16 weights); context window 512 to 514 tokens (2019-09-26).","keys":[{"file":"config","key":"max_position_embeddings","before":512,"after":514}],"consequence":{"params":{"before":124032768,"after":124034304},"weightsGbBf16":{"before":0.248,"after":0.248},"trainPeakGb":{"before":7.591,"after":7.611},"fingerprint":null,"context":{"before":512,"after":514},"eos":null,"bos":null},"fromSha":"bbbb96c1319c63786cbe53f5d35e4e857b4a0c95","toSha":"3094fd15f106304932bbeceb2c40ecd7e82ec6cd","commit":"https://huggingface.co/FacebookAI/roberta-base/commit/3094fd15f106304932bbeceb2c40ecd7e82ec6cd","source":"backfill"},{"id":"FacebookAI/roberta-large","date":"2019-09-26T17:03:17.000Z","classes":["structural"],"sentence":"roberta-large: derived parameters 354.3M to 354.3M (+2.0K, +0.00 GB of bf16 weights); context window 512 to 514 tokens (2019-09-26).","keys":[{"file":"config","key":"max_position_embeddings","before":512,"after":514}],"consequence":{"params":{"before":354255872,"after":354257920},"weightsGbBf16":{"before":0.709,"after":0.709},"trainPeakGb":{"before":20.564,"after":20.618},"fingerprint":null,"context":{"before":512,"after":514},"eos":null,"bos":null},"fromSha":"f0c702d5c384ae53c3a7f7eec945c70c93526a32","toSha":"5cae4475acf96517d6c1e33b4bf8f9444539988d","commit":"https://huggingface.co/FacebookAI/roberta-large/commit/5cae4475acf96517d6c1e33b4bf8f9444539988d","source":"backfill"},{"id":"FacebookAI/roberta-base","date":"2019-09-26T16:34:03.000Z","classes":["structural"],"sentence":"roberta-base: derived parameters 124.0M to 124.0M (-1.5K, -0.00 GB of bf16 weights); context window 514 to 512 tokens (2019-09-26).","keys":[{"file":"config","key":"max_position_embeddings","before":514,"after":512}],"consequence":{"params":{"before":124034304,"after":124032768},"weightsGbBf16":{"before":0.248,"after":0.248},"trainPeakGb":{"before":7.611,"after":7.591},"fingerprint":null,"context":{"before":514,"after":512},"eos":null,"bos":null},"fromSha":"02c245a6111bcb468a2e780ae76635895d9b8d66","toSha":"bbbb96c1319c63786cbe53f5d35e4e857b4a0c95","commit":"https://huggingface.co/FacebookAI/roberta-base/commit/bbbb96c1319c63786cbe53f5d35e4e857b4a0c95","source":"backfill"},{"id":"FacebookAI/roberta-large","date":"2019-09-26T16:34:03.000Z","classes":["structural"],"sentence":"roberta-large: derived parameters 354.3M to 354.3M (-2.0K, -0.00 GB of bf16 weights); context window 514 to 512 tokens (2019-09-26).","keys":[{"file":"config","key":"max_position_embeddings","before":514,"after":512}],"consequence":{"params":{"before":354257920,"after":354255872},"weightsGbBf16":{"before":0.709,"after":0.709},"trainPeakGb":{"before":20.618,"after":20.564},"fingerprint":null,"context":{"before":514,"after":512},"eos":null,"bos":null},"fromSha":"5d7774624540ab28a49a6155c2f97f021da08c23","toSha":"f0c702d5c384ae53c3a7f7eec945c70c93526a32","commit":"https://huggingface.co/FacebookAI/roberta-large/commit/f0c702d5c384ae53c3a7f7eec945c70c93526a32","source":"backfill"}]}