README.md

Reproduce ES with PARL

Based on PARL, we have implemented the Evolution Strategies (ES) algorithm and evaluate it in Mujoco environments. Its performance reaches the same level of indicators as the paper.

• ES in Evolution Strategies as a Scalable Alternative to Reinforcement Learning

Mujoco games introduction

Please see here to know more about Mujoco games.

Benchmark result

How to use

Dependencies

• Python3.7+
• paddlepaddle>=2.0.0
• parl>=2.1.1
• gym>=0.26.0
• mujoco>=2.2.2

Distributed training

To replicate the performance reported above, we encourage you to train with 24 or 48 CPUs.
If you haven't created a cluster before, enter the following command to create a cluster. For more information about the cluster, please refer to our documentation.

xparl start --port 8837 --cpu_num 24

Then we can start the distributed training by running:

python train.py

Training result will be saved in train_log with the training curve.

Reference

• Ray
• evolution-strategies-starter