search

Denys88 / rl_games

RL implementations
MIT License
820 stars 138 forks source link

BasePlayer modification #216

Closed romaf5 closed 1 year ago

romaf5 commented 1 year ago

Example:

config:
  env_config:
     ... train parameters ... 
  player:
      env_config:
         ... play parameters ...

Test plan:

  1. Train models for two different scenarios python runner.py --train --file rl_games/configs/brax/ppo_ant.yaml python runner.py --train --file rl_games/configs/ppo_cartpole.yaml

  2. Run model with & without proposed change. (No crash, similar results)

  3. Benchmark change (brax/ant)

Time Before After
real 1m32.151s 0m42.622s
user 1m21.269s 0m49.443s
sys 0m31.144s 0m17.270s