jmhb/VidDiffBench download history

jmhb/VidDiffBench is a video text to text dataset on the Hugging Face Hub. In the last 30 days it was downloaded 128 times (9 in the last 7 days), and 6,516 times in total. It ranks #95,127 among datasets by monthly downloads.

Dataset card for "VidDiffBench" This is the dataset / benchmark for Video Action Differencing (ICLR 2025), a new task that compares how an action is performed between two videos. This page introduces the task, the dataset structure, and how to access the data. See the paper for details on

Open jmhb/VidDiffBench on Hugging Face

Sister project: Paper Pulse, the upvote history of every Hugging Face Daily Paper.