https://github.com/semitable/seac
The official code base of Shared Experience Actor-Critic (NeurIPS2020)
https://github.com/semitable/seac
Last synced: 9 months ago
JSON representation
The official code base of Shared Experience Actor-Critic (NeurIPS2020)
- Host: GitHub
- URL: https://github.com/semitable/seac
- Owner: semitable
- Created: 2020-10-13T11:36:09.000Z (over 5 years ago)
- Default Branch: master
- Last Pushed: 2024-02-23T11:20:37.000Z (over 2 years ago)
- Last Synced: 2024-02-23T12:31:37.873Z (over 2 years ago)
- Language: Python
- Homepage:
- Size: 903 KB
- Stars: 14
- Watchers: 2
- Forks: 20
- Open Issues: 4
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# Shared Experience Actor Critic
This repository is the official implementation of [Shared Experience Actor Critic](https://arxiv.org/abs/2006.07169).
## Requirements
For the experiments in LBF and RWARE, please install from:
- [Level Based Foraging Official Repo](https://github.com/uoe-agents/lb-foraging)
- [Multi-Robot Warehouse Official Repo](https://github.com/uoe-agents/robotic-warehouse)
Also requires, PyTorch 1.6+
## Training - SEAC
To train the agents in the paper, navigate to the seac directory:
```
cd seac
```
And run:
```train
python train.py with
```
Valid environment configs are:
- `env_name=Foraging-15x15-3p-4f-v0 time_limit=25`
- ...
- `env_name=Foraging-12x12-2p-1f-v0 time_limit=25` or any other foraging environment size/configuration.
- `env_name=rware-tiny-2ag-v1 time_limit=500`
- `env_name=rware-tiny-4ag-v1 time_limit=500`
- ...
- `env_name=rware-tiny-2ag-hard-v1 time_limit=500` or any other rware environment size/configuration.
## Training - SEQL
To train the agents in the paper, navigate to the seac directory:
```
cd seql
```
And run the training script. Possible options are:
- `python lbf_train.py --env Foraging-12x12-2p-1f-v0`
- ...
- `python lbf_train.py --env Foraging-15x15-3p-4f-v0` or any other foraging environment size/configuration.
- `python rware_train.py --env "rware-tiny-2ag-v1"`
- ...
- `python rware_train.py --env "rware-tiny-4ag-v1"`or any other rware environment size/configuration.
## Evaluation/Visualization - SEAC
To load and render the pretrained models in SEAC, run in the seac directory
```eval
python evaluate.py
```
## Citation
```
@inproceedings{christianos2020shared,
title={Shared Experience Actor-Critic for Multi-Agent Reinforcement Learning},
author={Christianos, Filippos and Sch{\"a}fer, Lukas and Albrecht, Stefano V},
booktitle = {Advances in Neural Information Processing Systems},
year={2020}
}
```