Glint360K Face Recognition Dataset With 360K Identities

17,091,657 face images of 360,232 distinct individuals comprise this collection for training face recognition models. The dataset was cleaned and released by yayoimizuha, with the latest update recorded in October 2025. It is designed to support state-of-the-art performance on benchmarks like IJB-C and MegaFace.

Use Cases

• Train models to classify face images into one of 360,232 identity labels.
• Develop feature extraction models using the image data for face verification tasks.
• Benchmark model performance on large-scale test sets like IJB-C using the provided evaluation protocols.
• Investigate the effect of dataset scale on face recognition accuracy using over 17 million samples.

Strengths

• Contains 17,091,657 images, providing substantial scale for training.
• Covers 360,232 unique identities, offering high class diversity.

Limitations

• The specific geographic and demographic distribution of the 360,232 individuals is unknown, which may introduce bias.
• No information on image resolution, lighting conditions, or pose variation is provided to assess quality.

Provenance

Source

InsightFace Repository, cleaned and released by yayoimizuha.

Collection Method

Data was cleaned and merged from unspecified sources.

Freshness

Last updated on 2025-10-12.