Agent skill · nvidia
tao-generate-image-embeddings
Run TAO Data Services image embedding to turn a parquet of image filepaths into an embedding parquet using CLIP, SigLIP, or a TAO checkpoint. Use when a workflow needs embeddings before nearest-neighbor or unique-neighbor mining, or when the user asks to "embed images", "compute image embeddings", or "generate SigLIP embeddings".
What it needs
About 5k tokens when loaded.
What this skill does
TAO Generate Image Embeddings Use this skill to run TAO Data Services image embedding. The skill consumes a parquet of image filepaths and writes a parquet with an embedding column. Downstream mining skills (tao-mine-od-images, tao-mine-nearest-neighbors) consume its output. The container entrypoint is: Inputs The user can provide either an existing spec or the fields needed to generate one. Required spec fields: Field Meaning --- --- inputparquet Absolute path to a parquet containing image filepaths. outputparquet Absolute path where the embedding parquet is written. model CLIP or SigLIP. modelpath HuggingFace model id, local HF snapshot directory, or a TAO .pth/.ckpt checkpoint. Must match model. SigLIP: google/siglip-base-patch16-224 (768-dim, the template default). CLIP: openai/clip-vit-base-patch32 (512-dim). The validator rejects a recognizable model/path mismatch before launch. Common optional fields: Field Default Meaning --- ---: --- modelconfigpath "" TAO experiment spec path. Required only when modelpath is a TAO checkpoint. batchsize 64 Number of images processed in parallel. Lower it if the GPU runs out of memory. The input parquet must contain a filepath column. Any additional columns are carried through to the output verbatim, so metadata such as label survives into the embedding parquet. filepath values are transitive file inputs. When another workflow emits this spec, its platform action request must declare and mount every image root those paths reference; mounting only inputparquet is insufficient. Keep these mount-only folder inputs outside the native embedding YAML, because the Data Services config has no imageroot field. If the producing workflow freezes absolute paths, each embedding container must remount the corresponding source at the same compute-visible root. The default template is assets/defaultimageembeddings.yaml. …
How to use it
Reference it in AdaL, Claude Code, Cursor or any coding agent — nothing to install:
@skills nvidia/tao-generate-image-embeddings