search

PaddlePaddle / PARL

A high-performance distributed training framework for Reinforcement Learning
https://parl.readthedocs.io/
Apache License 2.0
3.22k stars 817 forks source link

fix bug of mappo #1060

Closed caoxixiya closed 1 year ago

caoxixiya commented 1 year ago

The redundant code in the official code has been deleted, new code can work well on cpu and gpu now.