Oct 27 6 days ago 12:45 Add task category to metadata nielsr HF Staff This PR adds the missing `task_categories` field (`question-answering`) to the YAML metadata of the dataset card, improving discoverability and alignment with Hub conventions. Existing metadata, configs, and README content are preserved. 02:13 Add arXiv link and citation Oct 25 8 days ago 22:08 Update arXiv badge 21:57 Update README 19:44 README: drop sample counts 19:44 README: wording 19:40 CASM summary.csv: pooled PSNR / SSIM / LPIPS only (as in the paper) 19:36 README: re-enable dataset viewer (casm / memobench configs) 19:35 README: disable dataset viewer 19:34 Drop VLM score columns from CASM summaries; README 19:34 README: pipeline + file structure 19:28 README: casm + memobench viewer configs 19:28 MemoBench results: worldkv 19:28 MemoBench results: reworld 19:27 MemoBench results: magicworld 19:27 MemoBench results: i3dm 19:27 MemoBench results: hydra 19:27 MemoBench results: framepack 19:11 README: single summary config (no auto train/test split) 19:11 CASM D results: worldkv 19:11 CASM D results: reworld 19:11 CASM D results: magicworld 19:11 CASM D results: i3dm 19:11 CASM D results: hydra 19:11 CASM D results: framepack 18:11 CASM: per-baseline summary.csv 18:10 CASM A/OV results: worldkv 18:10 CASM A/OV results: reworld 18:10 CASM A/OV results: magicworld 18:10 CASM A/OV results: i3dm 18:10 CASM A/OV results: hydra 18:10 CASM A/OV results: framepack 02:03 README: header badges + citation Oct 17 16 days ago 21:55 initial commit 
