Kinetics 700 Human Action Recognition Video Collection

Kinetics-700 is a large-scale collection of YouTube video URLs curated for human action recognition. The dataset was uploaded by atalaydenknalbant and was last updated on the Hugging Face platform in August 2025. It consists of 22 compressed archives that must be downloaded and decompressed to access the complete video collection.

Use Cases

• Train deep learning models like 3D CNNs or Transformers to classify human actions from video URLs.
• Benchmark action recognition model performance on a large-scale dataset of YouTube videos.
• Analyze the diversity and distribution of human actions represented in web-sourced video content.

Strengths

• Large scale collection of video data, as indicated by platform tags.
• Contains 22 separate archives, suggesting a substantial volume of video material.

Limitations

• Specific row count, video duration statistics, and class distribution are unknown.
• Potential for broken or removed YouTube URLs over time, a common issue with web-sourced video datasets.

Provenance

Source

atalaydenknalbant on Hugging Face, derived from the original Kinetics-700 research dataset.

Collection Method

Collection of YouTube video URLs, as stated in the description.

Freshness

Dataset card was last updated on 2025-08-21.

Geography

Platform tags suggest a US region focus, but specific geographic coverage is unknown.