Changes / Feeds / HuggingFaceTB/SmolVLM2-500M-Video-Instruct
HuggingFaceTB/SmolVLM2-500M-Video-Instruct
2 changes to a tracked key since 2025-02-11, newest first. Head as last read: 7b375e1b73b11138ff12fe22c8f2822d8fe03467.
Every tensor that differs from HuggingFaceTB/SmolVLM-500M-Instruct is also stored in another dtype (BF16 to F32): 489 of 489. A cast changes every byte, so the hashes cannot say whether values were also edited. Compared tensor by tensor from sha256 fingerprints: 7b375e1b73b1 against HuggingFaceTB/SmolVLM-500M-Instruct at a7da5b986cb5. A hash is equal or not; it says nothing about what an edit does.
| Date | Advisory | What changed, and what it does |
|---|---|---|
| 2025-02-16 | x_NXMC-2025-0178 breaking | layer count 32 declared (it was unset, so the library default for the class applied); derived parameters 185.3M to 332.9M only because our reconstruction stops falling back to its own default; whether the loaded model changed depends on the library default, which this record does not check (2025-02-16).2fb535e21a50 -> edf555f29eae Open in app |
| 2025-02-12 | x_NXMC-2025-0177 breaking | model_type "idefics3" to "smolvlm"; a different class is instantiated from the same id (2025-02-12).3d70c32f1d48 -> d52b9cddd529 Open in app |
Pin it, and let a bot move the pin
from_pretrained("HuggingFaceTB/SmolVLM2-500M-Video-Instruct", revision="7b375e1b73b11138ff12fe22c8f2822d8fe03467")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/HuggingFaceTB/SmolVLM2-500M-Video-Instruct.html)