MSVD (Microsoft Video Description)

1,058 videos (21.0 GB) with 680 subtitle files, downloaded at source quality and hosted for direct download.

Coverage: 1,058 of the 1,587 source video IDs (66.7%). 529 source videos were unavailable at fetch time (private, removed, members-only or geo-blocked) and are excluded.

Public release: capped at 1 TB (1,000,000,000,000 bytes), including media and indexes. This release retains 1,058 of the 1,058 previously hosted source videos. See selection.json for the selection method and byte counts. This is a selected subset, not the full upstream corpus.

What's inside

• metadata.jsonl — one row per video: video_id, video_url (direct HTTPS download, CDN-backed), subtitle_url, youtube_url, size_bytes
• The 1,058 MP4s (21.0 GB) and 680 subtitle files are hosted on Hugging Face storage and served via the video_url column
• captions.parquet — 59,025 academic annotations (source: MSVD_video_corpus.csv)

Quick start

from datasets import load_dataset

# videos (default config): one row per video with its direct download URL
ds = load_dataset("TornadoLabs/msvd", "videos", split="train")
print(ds[0])

# download one video
import requests
r = requests.get(ds[0]["video_url"])
open(f"{ds[0]['video_id']}.mp4", "wb").write(r.content)


# annotations for the published source videos
ann = load_dataset("TornadoLabs/msvd", "annotations", split="train")

How this dataset was built

Every video was fetched through the Tornado API bulk pipeline and delivered straight into Hugging Face storage via its S3-compatible gateway — the exact workflow documented here.

Of the original ID list, only videos still publicly available on YouTube are included (dead/private videos are listed upstream and excluded here).

Provenance & citation

• Original dataset: https://www.cs.utexas.edu/~ml/clamp/videoDescription/
• Paper: Chen & Dolan, ACL 2011 — Collecting Highly Parallel Data for Paraphrase Evaluation
• 1,970 clips YouTube annotés de ~122k descriptions en 35 langues (85k EN).

The video files remain the property of their respective YouTube uploaders. The annotations belong to the original dataset authors — cite their paper if you use them. This re-hosting is provided for reproducible research; if you are a rights holder and want content removed, open a discussion on this repo.

Request a larger or custom delivery

For larger volumes, different source IDs or a custom selection, request a delivery quote from Tornado API. Deliverable coverage depends on source availability and applicable rights. This is not a guarantee of the complete original corpus or a grant of commercial rights to the source videos. The upstream terms continue to apply.