KO | EN

GitLite

Minimalist Open Source Repository Search Engine

gitlite — search

#developer-tools #typescript #claude-code #codex #dsh #dsh-plugin #ai-agent #ai-agents #deepseek-harness #agent-skills #macos #cli

decentralized-multiarm

★ 165 Open GitHub ↗

[CoRL 2020] Learning a Decentralized Multiarm Motion Planner

// repository documentation

Was this content helpful?

★ 0 (0 ratings)

# Learning a Decentralized Multiarm Motion Planner

[Huy Ha](https://www.haquochuy.com),
[Jingxi Xu](https://jxu.ai/),
[Shuran Song](https://www.cs.columbia.edu/~shurans/),
<br>
Columbia University, New York, NY, United States<br>
[CoRL 2020](https://www.robot-learning.org/)

### [Project Page](https://multiarm.cs.columbia.edu/) | [Video](https://youtu.be/GNos793PFG4) | [Arxiv](https://arxiv.org/abs/2011.02608)

![](https://raw.githubusercontent.com/real-stanford/decentralized-multiarm/main/assets/teaser.gif)

Visualizations created using [PyBullet-Blender recorder](https://github.com/huy-ha/pybullet-blender-recorder)

## Setup

Python 3.7 dependencies:
 - PyTorch 1.6.0
 - pybullet
 - numpy
 - numpy-quaternion
 - ray
 - tensorboardX

We've prepared a conda YAML file which contains all the necessary dependencies. To use it, run

```sh
conda env create -f environment.yml
conda activate multiarm
```

## Evaluate the pretrained motion planner

In the repo's root, download the pretrained weights and evaluation benchmark
```sh
wget -qO- https://multiarm.cs.columbia.edu/downloads/checkpoints/ours.tar.xz | tar xvfJ -
wget -qO- https://multiarm.cs.columbia.edu/downloads/data/benchmark.tar.xz | tar xvfJ -
```
Then evaluate the pretrained weights on the benchmark in static mode with
```sh
python main.py --mode benchmark --tasks_path benchmark/ --load ours/ours.pth --num_processes 1 --gui
```
You can remove `--gui` to run headless, and use more CPU cores with `--num_processes 16`.

To summarize the benchmark results
```sh
python summary.py ours/benchmark_score.pkl
```
To evaluate the pretrained weights on the benchmark in dynamic mode, run 
```sh
python benchmark_dynamic.py --mode benchmark --tasks_path benchmark/ --load ours/ours.pth --num_processes 1 --gui
```

## Train a decentralized multi-arm motion planner

In the repo's root, download the training tasks and expert demonstration dataset

```sh
wget -qO- https://multiarm.cs.columbia.edu/downloads/data/tasks.tar.xz | tar xvfJ -
wget -qO- https://multiarm.cs.columbia.edu/downloads/data/expert.tar.xz | tar xvfJ -
```
Then train a decentralized multi-arm motion planner from scratch with 
```sh
mkdir runs
python main.py --config configs/default.json --tasks_path tasks/ --expert_waypoints expert/ --num_processes 16 --name multiarm_motion_planner
```

## Running the 6 DOF Bin Pick and Place Demo

See [demo/README.md](demo/README.md) for instructions.

## Citation

```
@inproceedings{ha2020multiarm,
	title={Learning a Decentralized Multi-arm Motion Planner},
	author={Ha, Huy and Xu, Jingxi and Song, Shuran},
	booktitle={Conference on Robotic Learning (CoRL)},
	year={2020}
}
```

decentralized-multiarm

Explore Similar Repositories

Was this content helpful?

Recent Feedback

Download README