Changes / Feeds / NousResearch/Hermes-3-Llama-3.1-8B
NousResearch/Hermes-3-Llama-3.1-8B
11 changes to a tracked key since 2024-07-28, newest first. Head as last read: 896ea440e5a9e6070e3d8a2774daf2b481ab425b.
| Date | Advisory | What changed, and what it does |
|---|---|---|
| 2024-09-08 | x_NXMC-2024-0160 | 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).489839286ef3 -> 896ea440e5a9 |
| 2024-09-08 | x_NXMC-2024-0159 breaking | tokenizer_config.json was removed; whatever defaults it set now come from the library (2024-09-08).5ba4922000a4 -> ebcf7c1aca87 |
| 2024-09-08 | x_NXMC-2024-0158 | 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).5e3fd4817e21 -> 5ba4922000a4 |
| 2024-09-05 | x_NXMC-2024-0157 | 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).aa50cfc2048c -> 5e3fd4817e21 |
| 2024-08-22 | x_NXMC-2024-0156 | 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).67bf4aca9f42 -> aa50cfc2048c |
| 2024-08-18 | x_NXMC-2024-0155 | 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).db6a44dd23ec -> f537e1048542 |
| 2024-08-06 | x_NXMC-2024-0154 breaking | the chat template changed and renders differently (system+user, multi-turn, tool-call); 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).248bb9b3e1f0 -> db6a44dd23ec |
| 2024-07-31 | x_NXMC-2024-0153 breaking | 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).5aada9f97327 -> 9949f486fbfa |
| 2024-07-31 | x_NXMC-2024-0152 | pad token "<|end_of_text|>" to "<|im_end|>"; batched training masks a different id (2024-07-31).34ccd9045c44 -> 5aada9f97327 |
| 2024-07-31 | x_NXMC-2024-0151 breaking | EOS changed from 128001 to 128040; anything fine-tuned or served against the old value stops at a different token (2024-07-31).d4ed47f114cf -> 34ccd9045c44 |
| 2024-07-31 | x_NXMC-2024-0150 breaking | EOS changed from 128001 to 128040; anything fine-tuned or served against the old value stops at a different token (2024-07-31).46ed7f46fd64 -> d4ed47f114cf |
Pin it, and let a bot move the pin
from_pretrained("NousResearch/Hermes-3-Llama-3.1-8B", revision="896ea440e5a9e6070e3d8a2774daf2b481ab425b")With the Renovate preset, a pull request moves this pin when the head moves, and links back here.
[](https://www.neurarch.com/changes-feeds/models/NousResearch/Hermes-3-Llama-3.1-8B.html)