Agent skill · nvidia
tao-finetune-video-clip
InternVideo2-CLIP L14 (TAO video_clip) for video-text retrieval, zero-shot classification, embedding extraction, LoRA fine-tuning, ONNX export, and TensorRT deployment. Use when the user asks to "fine-tune IV2CLIP", "run video_clip train/evaluate/inference/export", "build a Video-CLIP TensorRT engine", "InternVideo2-CLIP on KPI chunks", or "TAO video_clip on vadr1_chunks JSON".
What it needs
About 6k tokens when loaded.
What this skill does
InternVideo2-CLIP (TAO videoclip) Standalone install? If this session was not initialized by the TAO skill bank plugin, run the tao-setup skill first (host preflight, credentials, cross-skill discovery). TAO task videoclip wraps OpenGVLab InternVideo2-CLIP L14. The PyTorch image provides train, evaluate, inference, export, and defaultspecs. TAO Deploy provides gentrtengine, TensorRT evaluate, and TensorRT inference. Container images and per-action commands are in references/skillinfo.yaml and references/tao-deploy-video-clip.skillinfo.yaml. Starting specs are in references/spectemplate.yaml. Release note: The pinned PyTorch image is the TAO 7.2 release-candidate build validated for Video-CLIP. It includes PyAV 17.1.0 as the primary decoder and ONNXScript 0.7.1 for export, with decord absent. The TAO Deploy image is pinned independently because gentrtengine and TensorRT-backed actions do not run in the PyTorch image. Known-broken images: interim builds cut before tao-pytorch commit 0cc31de4 ship a videoclip package with no model.backbones submodule, so train/evaluate/inference die at import while videoclip --help still exits 0. Images without PyAV also fail at data loading. Run both import checks in the preflight below before pulling data or launching a run. Train Action Policy AutoML is not packaged for this model skill. Always use direct videoclip actions even when a higher-level request mentions AutoML. Non-train actions stay in this skill. Quick Start (local Docker) Use the pinned TAO container declared in references/skillinfo.yaml. Pull with NGCKEY when the image is not cached locally. Expected workspace layout (host paths bind-mounted into the container): Docker options for all actions (skill-eval CI uses the same $WORKSPACEDIR bind-mount pattern): Preflight (host): Train: Evaluate: Inference: Export: TensorRT Deploy Use the independently pinned TAO Deploy image after PyTorch export. …
How to use it
Reference it in AdaL, Claude Code, Cursor or any coding agent — nothing to install:
@skills nvidia/tao-finetune-video-clip