datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
MMOU-Videos
MMOU Videos
This repository solely hosts the raw video files and captions for the MMOU benchmark. For the actual dataset, questions, evaluation scripts, and full documentation, please visit nvidia/MMOU.
What This Repository Contains
This repository is the storage companion for MMOU. It contains the heavy media assets needed to run the benchmark:
Videos: 9,049 root-level real-world videos stored as MP4 files and sampled at 720p.
Captions… See the full description on the dataset page: https://huggingface.co/datasets/sonalkum/MMOU-Videos.remake_videosXMER-VideosBandicam-Videos-And-Othersvideo-storagevideos-promptsiemocap-with-videosvideos-testDiary of a CEO ~10 hours, took ~15 minutes on tiny, ~5 minutes on EC2
MELD-videos-absolute-pathsvideosdeclassicoting-videosVideo_Speech_Actor_24Ditto_videos_hdtf_400_audio_60s_chunks_all_preprocessyoutube_videoscrema_d_Fusion_Test_Videosvideo_samplesvideos-test2video_screening_voice