sapiens2

Sapiens2

Sapiens2 is a family of high-resolution vision transformers pretrained on 1 billion human images — designed for human-centric tasks such as pose estimation, body-part segmentation, surface normals, pointmaps, and human matting.

This is the index repository: each variant lives in its own model repo (linked below).

• 📄 Paper: arXiv:2604.21681
• 🌐 Project Page: rawalkhirodkar.github.io/sapiens2
• 💻 Code: github.com/facebookresearch/sapiens2
• 📚 Collection: Sapiens2 on HuggingFace

Pretrained Backbones

Task Checkpoints

Pose Estimation

Body-Part Segmentation

Surface Normal Estimation

Pointmap Estimation

Human Matting

License

Released under the Sapiens2 License.

Citation

@article{khirodkarsapiens2,
  title={Sapiens2},
  author={Khirodkar, Rawal and Wen, He and Martinez, Julieta and Dong, Yuan and Su, Zhaoen and Saito, Shunsuke},
  journal={arXiv preprint arXiv:2604.21681},
  year={2026}
}