Changes / Feeds / OpenGVLab/InternVL2_5-4B
OpenGVLab/InternVL2_5-4B
4 changes to a tracked key since 2024-11-21, newest first. Head as last read: 2cf4a8158bbc40d35015e7c63b527890de4d27b3.
GPU memory to fine-tune it, card by card: InternVL2_5-4B on /m/.
| Date | Advisory | What changed, and what it does |
|---|---|---|
| 2025-09-11 | x_NXMC-2025-0030 breaking | the chat template changed and renders differently (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 (2025-09-11).de2c1f985d9a -> 2cf4a8158bbc |
| 2025-09-11 | x_NXMC-2025-0029 breaking | the chat template changed and renders differently (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 (2025-09-11).f01b74c11c53 -> de2c1f985d9a |
| 2024-11-29 | x_NXMC-2024-0021 breaking | 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).74b297953c78 -> f01b74c11c53 |
| 2024-11-26 | x_NXMC-2024-0020 breaking | 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).e21cfb8cc876 -> 74b297953c78 |
Pin it, and let a bot move the pin
from_pretrained("OpenGVLab/InternVL2_5-4B", revision="2cf4a8158bbc40d35015e7c63b527890de4d27b3")
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/OpenGVLab/InternVL2_5-4B.html)