Verified record · from the answer index
tao-generate-video-reasoning-annotations
skilltested · static-verifiedscore 46/100signed passport ✓1,618 uses
Multi-step video annotation pipeline that turns raw videos into Chain-of-Thought training data — multi-level captions, structured descriptions, and QA pairs (MCQ, binary, open-ended) with reasoning traces, via VLM/LLM distillation. Use when
Declared tools (1)
tao-generate-video-reasoning-annotations
skill:nvidia/skills/tao-generate-video-reasoning-annotations · published by nvidia · source repository