TikTok-10M: 10 Million Short-Form Video Posts for Multimodal Analysis

TikTok-10M is a large-scale dataset containing 10 million short-form posts from TikTok, curated to bridge the gap between academic video datasets and actual user-generated content. It was created by author Nikioooo and last updated on Hugging Face in April 2026. The dataset is designed to provide researchers with authentic patterns and characteristics of modern short-form video content.

Use Cases

• Train video understanding models based on authentic short-form video content.
• Conduct multimodal learning research based on the combination of video, image, and potentially text data.
• Analyze social media content patterns and trends based on characteristics of modern user-generated videos.
• Benchmark algorithms for short-form video recommendation or classification based on large-scale, real-world data.

Strengths

• Contains 10 million data points, indicating a large-scale collection.
• Designed to provide authentic patterns from real user-generated content, bridging a research gap.
• Last updated on 2026-04-01, suggesting recent maintenance.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is confirmed but specific file formats, size, and license details are unknown.
• Data may reflect geographic or temporal bias inherent to its source platform.

Provenance

Source

Hugging Face, uploaded by author Nikioooo.

Collection Method

Curated from TikTok; specific gathering method is not detailed in the provided description.

Time Range

null

Freshness

Last updated 2026-04-01 03:16:07.

Geography

null