VFHQ 25FPS: Face Video and Audio Dataset

VFHQ 25FPS is a dataset containing video and audio data, organized into a 'data' folder. The dataset was uploaded to Hugging Face by Holasyb918 and was last updated on 2026-01-26. It follows the same license as the original VFHQ dataset, and users are advised to ensure they have the necessary permissions to use the underlying VFHQ data.

Use Cases

• Train face video generation models based on the video data.
• Develop video editing techniques for face videos based on the provided video content.
• Create multimodal AI systems that synchronize video and audio based on the dataset's structure.
• Benchmark face video quality or temporal consistency based on the 25 frames-per-second video data.

Strengths

• Dataset structure is explicitly described with a dedicated 'data' folder for video and audio.
• Last update timestamp (2026-01-26) is provided, indicating recent maintenance.

Limitations

• Description metadata is limited; actual data quality, resolution, and content require manual inspection after download.
• Row count, file formats, and column-level documentation are absent, which may limit suitability assessment.
• The license is described as following the original VFHQ license, but the specific terms are not detailed here.

Provenance

Source

Holasyb918 on Hugging Face, derived from the original VFHQ dataset.

Freshness

Last updated 2026-01-26 08:42:38; freshness should be verified.