Add README.md
Browse files
README.md
ADDED
|
@@ -0,0 +1,16 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: apache-2.0
|
| 3 |
+
library_name: pytorch
|
| 4 |
+
pipeline_tag: image-to-video
|
| 5 |
+
---
|
| 6 |
+
# HandWorld
|
| 7 |
+
|
| 8 |
+
Weights for **HandWorld: Hand-Centric Unified Video Action Generation**, CVPR 2026.
|
| 9 |
+
|
| 10 |
+
[Paper](https://openaccess.thecvf.com/content/CVPR2026/html/Sun_HandWorld_Hand-Centric_Unified_Video_Action_Generation_CVPR_2026_paper.html)
|
| 11 |
+
|
| 12 |
+
- `stage2.safetensors`: complete action-conditioned video model.
|
| 13 |
+
- `stage3.safetensors`: complete joint video/action prediction model.
|
| 14 |
+
- `preprocess/detector.pt`: optional left/right hand detector for RGB preprocessing.
|
| 15 |
+
|
| 16 |
+
Place the desired stage file in the inference repository's `weights/` directory. Download `models_t5_umt5-xxl-enc-bf16.pth`, `Wan2.2_VAE.pth`, and `google/umt5-xxl/` separately from the official [Wan-AI/Wan2.2-TI2V-5B](https://huggingface.co/Wan-AI/Wan2.2-TI2V-5B/tree/main) repository into `weights/Wan2.2-TI2V-5B/`, preserving those relative paths. No earlier-stage checkpoint or original Wan video-DiT shards are required. HaMeR and licensed MANO assets are obtained separately for preprocessing.
|