Collections

Discover the best community collections!

Collections trending this week
ViDoRe Chunk OCR (baseline)
The ViDoRe benchmark was passed to Unstructured to partition each page into text chunks. Detected figures/tables were captioned with Claude 3-Sonnet.