KO | EN

GitLite

Minimalist Open Source Repository Search Engine

gitlite — search

#dsh #dsh-plugin #typescript #ai-agent #developer-tools #claude-code #codex #deepseek-harness #ai-agents #react #mcp #rust

CEDex-Grasp

★ 32 Open GitHub ↗

Official repository of CEDex: Cross-Embodiment Dexterous Grasp Generation at Scale from Human-like Contact Representations (ICRA 2026)

// repository documentation

Was this content helpful?

★ 0 (0 ratings)

# CEDex Code and Dataset
Official Repository of **CEDex: Cross-Embodiment Dexterous Grasp Generation at Scale from Human-like Contact Representations (ICRA 2026)**

<a href="https://georgewuzy.github.io/" style="color: #4A90E2 !important;">Zhiyuan Wu</a>1,


 <a href="https://rolpotamias.github.io/" style="color: #4A90E2 !important;">Rolandos Alexandros Potamias</a>2,


 <a href="https://scholar.google.com/citations?user=osQ5dAkAAAAJ&hl=zh-CN&oi=ao" style="color: #4A90E2 !important;">Xuyang Zhang</a>1,


 <a href="https://zhongqunzhang.github.io/" style="color: #4A90E2 !important;">Zhongqun Zhang</a>3,


 <a href="https://jiankangdeng.github.io/" style="color: #4A90E2 !important;">Jiankang Deng</a>2,


 <a href="https://shanluo.github.io/" style="color: #4A90E2 !important;">Shan Luo</a>1

1 King's College London, 2 Imperial College London, 3 Nankai University

## Overview
![Pipeline](https://raw.githubusercontent.com/GeorgeWuzy/CEDex-Grasp/main/assets/pipeline.gif)  
In this paper, we propose **CEDex**, a novel cross-embodiment dexterous grasp synthesis method that bridges human grasping kinematics and robot kinematics by aligning robot kinematic models with generated human-like contact representations. Using CEDex, we construct the largest cross-embodiment grasp dataset to date, comprising **500K objects** across four gripper types with **20M total grasps**.

## Updates
- [May. 2026] We have released **more synthesis object set** grasp data.

- [Apr. 2026] We have added training code and configuration for CVAE model (based on [ContactGen](https://github.com/stevenlsw/contactgen)).

- [Feb. 2026] We have released our **grasp generation code**.

- [Oct. 2025] We have released a part of our **synthesis object set** grasp data, including objects from [Objaverse](https://objaverse.allenai.org/). We will soon release more data. Please stay tuned for updates!

![Simulation Objects](https://raw.githubusercontent.com/GeorgeWuzy/CEDex-Grasp/main/assets/sim_objects.gif)

- [Oct. 2025] We have released the **real-world object set** grasp data, including objects from [ContactDB](https://contactdb.cc.gatech.edu/) and [YCB](https://www.ycbbenchmarks.com/).

![Real-World Objects](https://raw.githubusercontent.com/GeorgeWuzy/CEDex-Grasp/main/assets/rw_objects.gif)

## Dependencies

Create a new conda environment. My CUDA version (nvcc --version) is 12.4
```bash  
conda create -n cedex python=3.8  
conda activate cedex
pip install torch==2.4.0 torchvision==0.19.0 torchaudio==2.4.0 --index-url https://download.pytorch.org/whl/cu124
git clone https://github.com/facebookresearch/pytorch3d.git
cd pytorch3d
# export MAX_JOBS=4 # set MAX_JOBS if your workstation doesn't have enough memory
python setup.py install
unset MAX_JOBS
cd ../
```

Install required packages.
```bash  
pip install -r requirements.txt
```

Install Isaac Gym Environment (for validation).

Download [Isaac Gym](https://developer.nvidia.com/isaac-gym/download) from the official website, then:
```bash  
tar -xvf IsaacGym_Preview_4_Package.tar.gz
cd isaacgym/python
pip install -e .
```

Install pointnet_lib for ContactGen.
```bash
cd contactgen
cd pointnet_lib
python setup.py install
cd ../../
```

## Data Preparation

For **real-world object set** grasp data, download the robot and object data from [OneDrive link](https://emckclac-my.sharepoint.com/:u:/g/personal/k24021603_kcl_ac_uk/IQDDknATihQyQJ9hfHt7_5OBAY9nqAcSQUMLXlCe34nohU0?e=fRTWMP) and extract the contents to the `data/` directory. Our grasp data is in `cedex/{robot_name}.pt`.

For **synthesis object set** grasp data, download the grasp data from [OneDrive link](https://emckclac-my.sharepoint.com/:f:/g/personal/k24021603_kcl_ac_uk/IgA3UfGGDSm4QJrkVD1TCENJAeumUO37dIUK1SCecGD6auE) and object data from [OneDrive link](https://emckclac-my.sharepoint.com/:u:/g/personal/k24021603_kcl_ac_uk/IQD_44y-unfpSoVEDIxRs1lkAXndXxC1DnI5X1tR8YN-tlg?e=IulJLh), and put the contents to the `cedex_objaverse/` directory. `{robot_name}_isaac.pt` refer to grasp data after grasp execution in Isaac with no penetration.

## Usage

### Data Usage

In our data, each grasp data contains the following keys: `['q']` and `['object_name']`. You can use the dataset by calling the robot hand through the `utils_model/HandModel.py`. This implementation is consistent with [GenDexGrasp](https://github.com/tengyu-liu/GenDexGrasp), but please note that there are differences in object URDF files. Make sure to use our provided objects.

To visualize the grasping data, you can use the following command:

```bash  
python vis_cedex.py --input_file cedex/allegro.pt  
python vis_cedex.py --input_file cedex_objaverse/shadowhand.pt  
```

### Grasp Validation

For grasp validation, our implementation is based on [DRO-Grasp](https://github.com/zhenyuwei2003/DRO-Grasp). We have modified some parameters of the controller. You can evaluate the grasping performance using the following command:

```bash
python eval_grasp.py --input_file cedex/allegro.pt --object_name ycb+055_baseball # --use_gui --eval_diversity
```

### Grasp Generation

You can use CEDex to generate your own grasp data.

**Stage 0** (Optional): Train CVAE Model for Human-Like Contact Map Generation (based on [ContactGen](https://github.com/stevenlsw/contactgen)). We recommend to finetune on the pretrained checkpoint. 
Download human grasp data from [OneDrive link](https://emckclac-my.sharepoint.com/:f:/g/personal/k24021603_kcl_ac_uk/IgDoRT6L-YGfSpZ1FisgCUHQAUQJu7Ls8cqIXskxWOBQex0) and put it to `data/`
```bash
cd contactgen
python train.py
cd ../
```

**Stage 1**: Generate Human-Like Contact Maps
```bash
cd contactgen
python inf_contactdb.py
python inf_ycb.py
cd ../
```

**Stage 2**: Generate Grasps
```bash
python generate_data.py --robot_name barrett --dataset contactdb
```
**Note**: Some hyperparameters can be adjusted to get more stable grasp results or more diverse grasp results. 
1) Initial positions of robot hands in `utils_model/CMapAdam.py`
2) Hyperparameters of physical constraints in `utils_model/CMapAdam.py`
3) Sort the top-k saved grasps in `generate_grasp.py` by different properties in contact energy

**Stage 3**: Visualize Generated Grasps
```bash
python vis_generated_grasp.py --robot_name barrett --input_dir logs/dataset_generation_20260202_162359/barrett_contactdb/ # replace the input_dir to your log dir. 
```

**Stage 4**: Validate and Filter Generated Grasps
```bash
python eval_grasp_filtered.py --robot_name barrett --logs_path logs/dataset_generation_20260202_162359/barrett_contactdb/ # --use_gui # replace the logs_path to your log dir
```

### Adapt to New Hands

Our CEDex currently supports 1. ShadowHand, 2. Allegro, 3. Barret, 4. Robotiq-3F, 5. Leap Hand. To adapt to new hands, you need to :

1. Build the human-robot remapping in `def remap_part_labels` in `utils_model/HandModel.py`. We have provided `visualize_hand_parts` command in `utils_model/HandModel.py` for reference. 
2. Define `self.dis_key_point` and `self.keypoints` in `utils_model/HandModel.py`. We have provided `visualize_hand_surface_and_keypoints` command in `utils_model.HandModel.py` for reference. To define the keypoints, please refer to this [issue](https://github.com/4DVLab/DexGrasp-Anything/issues/9).

## Citation

If you find this work helpful, please consider citing us using the following BibTeX entry:

```bibtex  
@article{wu2025cedex,  
  title={CEDex: Cross-Embodiment Dexterous Grasp Generation at Scale from Human-like Contact Representations},  
  author={Wu, Zhiyuan and Potamias, Rolandos Alexandros and Zhang, Xuyang and Zhang, Zhongqun and Deng, Jiankang and Luo, Shan},  
  journal={arXiv preprint arXiv:2509.24661},  
  year={2025}  
}  
```

## Contact

If you have any questions, feel free to contact me through email at [zhiyuan.1.wu@kcl.ac.uk](mailto:zhiyuan.1.wu@kcl.ac.uk).

CEDex-Grasp

Explore Similar Repositories

Was this content helpful?

Recent Feedback

Download README