The Dataset Viewer has been disabled on this dataset.

FDABench-File

Raw multimodal corpus distributed alongside the FDABench benchmark.

Contains 1,418 PDFs, 246 images, 108 audio clips, 81 video clips, plus transcripts and a manifest, organized into 50 domain categories.

Use huggingface_hub.snapshot_download("FDAbench2026/FDABench-File", repo_type="dataset") to fetch the full tree, or hf_hub_download(...) for individual files.

Downloads last month
91