zhenyuxie-zhzh/hot3d-minimax

HOT3D EgoHOI data + Stage-1 caches for the MiniMax-H3 EgoHOI fine-tune Everything the MiniMax-H3 EgoHOI LoRA + conditioner training (branch claude/minimax-h3-v2-pose of the project repository, commit 0d7fad8 or later) reads, so that a collaborator can start the pose-v2 experiment without re-deriving anything: the HOT3D EgoHOI clips, the canonical manifest, and the frozen-feature Stage-1 caches computed with the pinned MiniMax-H3 weights - the very caches the AMD-cluster training… See the full description on the dataset page: https://huggingface.co/datasets/zhenyuxie-zhzh/hot3d-minimax.

<iframe src="https://silytics.si/embed/item/230f40f2-5951-49b2-92ae-242923029d69" width="420" height="190" style="border:0" loading="lazy" title="Silytics entry"></iframe>

Silytics shows public metadata from Hugging Face and links to the original. Information may be incomplete or out of date. Last refreshed 18 October 2026. Licenses and terms of the original apply to the work itself. Not financial, legal or professional advice.