Dataset Viewer
Auto-converted to Parquet Duplicate
The dataset viewer is not available for this split.
The size of the content of the first rows (547062 B) exceeds the maximum supported size (200000 B) even after truncation. Please report the issue.
Error code:   TooBigContentError

Need help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.

VideoMind-HiREST-3fps-480-train

Unofficial, unmodified download subset of yeliudev/VideoMind-Dataset.

HiREST training annotations plus BOTH parts of the processed video archive. The source describes this version as 3 FPS, 480p, no audio. Video archive contents are unmodified and have NOT been filtered to training video IDs. Download both parts; they form one split archive and are not independently extractable. Concatenate in .00, .01 order and extract in the intended processing environment. Align annotation time/frame coordinates to the source convention, not merely to the compressed media FPS.

Source revision: 3518d8e8c2c7c1ceec4b242afc07162bbafa33d8. Data size: 85,095,541,180 bytes.

File Bytes
hirest/all_data_train.json 514,671
hirest/videos_3fps_480_noaudio.tar.gz.00 48,318,382,080
hirest/videos_3fps_480_noaudio.tar.gz.01 36,776,644,429

Download the whole repository. Original relative file paths are preserved. See source_manifest.json for file sizes and SHA256 hashes.

The original source card is preserved as SOURCE_README.md. License metadata, when present, is inherited from the source card; underlying media and annotations retain their applicable original terms. This packaging grants no additional rights and is not affiliated with the original authors.

Downloads last month
1,334