Upload README.md with huggingface_hub
Browse files
README.md
ADDED
|
@@ -0,0 +1,85 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: other
|
| 3 |
+
license_name: sapiens2-license
|
| 4 |
+
license_link: https://github.com/facebookresearch/sapiens2/blob/main/LICENSE.md
|
| 5 |
+
library_name: sapiens
|
| 6 |
+
tags:
|
| 7 |
+
- sapiens
|
| 8 |
+
- sapiens2
|
| 9 |
+
- human-centric
|
| 10 |
+
- vision-transformer
|
| 11 |
+
---
|
| 12 |
+
|
| 13 |
+
# Sapiens2
|
| 14 |
+
|
| 15 |
+
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, and pointmaps.
|
| 16 |
+
|
| 17 |
+
This is the **index** repository: each variant lives in its own model repo (linked below).
|
| 18 |
+
|
| 19 |
+
- 📄 **Paper:** [OpenReview (ICLR 2026)](https://openreview.net/pdf?id=IVAlYCqdvW)
|
| 20 |
+
- 🌐 **Project Page:** [rawalkhirodkar.github.io/sapiens2](https://rawalkhirodkar.github.io/sapiens2)
|
| 21 |
+
- 💻 **Code:** [github.com/facebookresearch/sapiens2](https://github.com/facebookresearch/sapiens2)
|
| 22 |
+
- 📚 **Collection:** [Sapiens2 on HuggingFace](https://huggingface.co/collections/facebook/sapiens2)
|
| 23 |
+
|
| 24 |
+
## Pretrained Backbones
|
| 25 |
+
|
| 26 |
+
| Model | Params | Repository |
|
| 27 |
+
|-------|--------|------------|
|
| 28 |
+
| Sapiens2-0.1B | 0.114 B | [facebook/sapiens2-pretrain-0.1b](https://huggingface.co/facebook/sapiens2-pretrain-0.1b) |
|
| 29 |
+
| Sapiens2-0.4B | 0.398 B | [facebook/sapiens2-pretrain-0.4b](https://huggingface.co/facebook/sapiens2-pretrain-0.4b) |
|
| 30 |
+
| Sapiens2-0.8B | 0.818 B | [facebook/sapiens2-pretrain-0.8b](https://huggingface.co/facebook/sapiens2-pretrain-0.8b) |
|
| 31 |
+
| Sapiens2-1B | 1.462 B | [facebook/sapiens2-pretrain-1b](https://huggingface.co/facebook/sapiens2-pretrain-1b) |
|
| 32 |
+
| Sapiens2-5B | 5.071 B | [facebook/sapiens2-pretrain-5b](https://huggingface.co/facebook/sapiens2-pretrain-5b) |
|
| 33 |
+
|
| 34 |
+
## Task Checkpoints
|
| 35 |
+
|
| 36 |
+
### Pose Estimation
|
| 37 |
+
|
| 38 |
+
| Model | Repository |
|
| 39 |
+
|-------|------------|
|
| 40 |
+
| Sapiens2-0.4B | [facebook/sapiens2-pose-0.4b](https://huggingface.co/facebook/sapiens2-pose-0.4b) |
|
| 41 |
+
| Sapiens2-0.8B | [facebook/sapiens2-pose-0.8b](https://huggingface.co/facebook/sapiens2-pose-0.8b) |
|
| 42 |
+
| Sapiens2-1B | [facebook/sapiens2-pose-1b](https://huggingface.co/facebook/sapiens2-pose-1b) |
|
| 43 |
+
| Sapiens2-5B | [facebook/sapiens2-pose-5b](https://huggingface.co/facebook/sapiens2-pose-5b) |
|
| 44 |
+
|
| 45 |
+
### Body-Part Segmentation
|
| 46 |
+
|
| 47 |
+
| Model | Repository |
|
| 48 |
+
|-------|------------|
|
| 49 |
+
| Sapiens2-0.4B | [facebook/sapiens2-seg-0.4b](https://huggingface.co/facebook/sapiens2-seg-0.4b) |
|
| 50 |
+
| Sapiens2-0.8B | [facebook/sapiens2-seg-0.8b](https://huggingface.co/facebook/sapiens2-seg-0.8b) |
|
| 51 |
+
| Sapiens2-1B | [facebook/sapiens2-seg-1b](https://huggingface.co/facebook/sapiens2-seg-1b) |
|
| 52 |
+
| Sapiens2-5B | [facebook/sapiens2-seg-5b](https://huggingface.co/facebook/sapiens2-seg-5b) |
|
| 53 |
+
|
| 54 |
+
### Surface Normal Estimation
|
| 55 |
+
|
| 56 |
+
| Model | Repository |
|
| 57 |
+
|-------|------------|
|
| 58 |
+
| Sapiens2-0.4B | [facebook/sapiens2-normal-0.4b](https://huggingface.co/facebook/sapiens2-normal-0.4b) |
|
| 59 |
+
| Sapiens2-0.8B | [facebook/sapiens2-normal-0.8b](https://huggingface.co/facebook/sapiens2-normal-0.8b) |
|
| 60 |
+
| Sapiens2-1B | [facebook/sapiens2-normal-1b](https://huggingface.co/facebook/sapiens2-normal-1b) |
|
| 61 |
+
| Sapiens2-5B | [facebook/sapiens2-normal-5b](https://huggingface.co/facebook/sapiens2-normal-5b) |
|
| 62 |
+
|
| 63 |
+
### Pointmap Estimation
|
| 64 |
+
|
| 65 |
+
| Model | Repository |
|
| 66 |
+
|-------|------------|
|
| 67 |
+
| Sapiens2-0.4B | [facebook/sapiens2-pointmap-0.4b](https://huggingface.co/facebook/sapiens2-pointmap-0.4b) |
|
| 68 |
+
| Sapiens2-0.8B | [facebook/sapiens2-pointmap-0.8b](https://huggingface.co/facebook/sapiens2-pointmap-0.8b) |
|
| 69 |
+
| Sapiens2-1B | [facebook/sapiens2-pointmap-1b](https://huggingface.co/facebook/sapiens2-pointmap-1b) |
|
| 70 |
+
| Sapiens2-5B | [facebook/sapiens2-pointmap-5b](https://huggingface.co/facebook/sapiens2-pointmap-5b) |
|
| 71 |
+
|
| 72 |
+
## License
|
| 73 |
+
|
| 74 |
+
Released under the [Sapiens2 License](https://github.com/facebookresearch/sapiens2/blob/main/LICENSE.md).
|
| 75 |
+
|
| 76 |
+
## Citation
|
| 77 |
+
|
| 78 |
+
```bibtex
|
| 79 |
+
@inproceedings{khirodkar2026sapiens2,
|
| 80 |
+
title={Sapiens2},
|
| 81 |
+
author={Khirodkar, Rawal and Wen, He and Martinez, Julieta and Dong, Yuan and Zhaoen, Su and Saito, Shunsuke},
|
| 82 |
+
booktitle={International Conference on Learning Representations (ICLR)},
|
| 83 |
+
year={2026}
|
| 84 |
+
}
|
| 85 |
+
```
|