MultiID-2M: Large-Scale Paired Dataset for Multi-Person Image Generation

MultiID-2M is a large-scale paired dataset constructed for multi-person scenarios in identity-consistent image generation. It provides diverse references for each identity to support the development of advanced diffusion-based models like WithAnyone. The dataset was created by author vcbe123 and was last updated on January 20, 2026.

Use Cases

• Training identity-consistent image generation models based on the dataset's paired multi-person references.
• Improving controllability over pose and expression in generated images based on the described features.
• Mitigating 'copy-paste' artifacts in multi-person image synthesis based on the dataset's construction purpose.

Strengths

• Designed specifically for multi-person scenarios, a focused use case.
• Provides diverse references for each identity, as stated in the description.
• Associated with the WithAnyone paper, suggesting a research foundation.

Limitations

• Description metadata is limited; actual data quality requires manual inspection after download.
• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.

Provenance

Source

vcbe123 on Hugging Face

Collection Method

Likely constructed for research on identity-consistent image generation, as described.

Freshness

Last updated 2026-01-20 12:32:56; freshness should be verified.