Data Enrichment#
Take a raw dataset with no metadata or labels and enrich it with image properties, CLIP embeddings, and a UMAP visualization — ready for semantic search, duplicate detection, and downstream labeling workflows.
Requirements#
Usage#
Ask your AI assistant:
"Enrich my dataset"
"Compute embeddings for my images"
"Set up my dataset for semantic search"
"I have raw images with no metadata, where do I start?"
The skill audits existing fields first and skips any step already done. It then computes image metadata, CLIP embeddings, and a UMAP visualization as sequential delegated background jobs — each step waits for the previous one to finish before starting.