Dataset · Hugging Face

facebook/action100m-preview

Name: action100m-preview
Creator: facebook
License: fair-noncommercial-research-license
Keywords: task_categories:video-classification, task_categories:video-text-to-text, language:en, license:fair-noncommercial-research-license, size_categories:100K<n<1M, format:parquet, modality:text, modality:video, library:datasets, library:dask, library:polars, library:mlcroissant, arxiv:2601.10592, region:us, video

Action100M: A Large-scale Video Action Dataset Paper | GitHub Action100M is a large-scale dataset constructed from 1.2M Internet instructional videos (14.6 years of duration), yielding ~100 million temporally localized segments with open-vocabulary action supervision and rich captions. It serves as a foundation for scalable research in video understanding and world modeling. Load Action100M Annotations Our data can be loaded from the 🤗 huggingface repo at… See the full description on the dataset page: https://huggingface.co/datasets/facebook/action100m-preview.

View on Hugging FaceReport an issue

Author / organization	facebook
License (as declared)	fair-noncommercial-research-license
Downloads	951 (as reported by Hugging Face)
Likes	151
Created	14 January 2026
Last updated at source	29 January 2026

videoaction

Embed this entry

<iframe src="https://silytics.si/embed/item/fb864b6c-382b-44ae-9bc7-56087010570f" width="420" height="190" style="border:0" loading="lazy" title="Silytics entry"></iframe>

Free, no tracking. More about feeds and embeds.

More from facebook

facebook/uco3dDataset
facebook/seamless-interactionDataset
facebook/anliDataset
facebook/multilingual_librispeechDataset
facebook/belebeleDataset
facebook/xnliDataset

Silytics shows public metadata from Hugging Face and links to the original. Information may be incomplete or out of date. Last refreshed 16 October 2026. Licenses and terms of the original apply to the work itself. Not financial, legal or professional advice.