agentos-project / agentos

The Python Component System (PCS) is an API and CLI for building, running, and sharing Python code. AgentOS is a set of libraries built on top of PCS that make it easy to build, run, and share agents that use Reinforcement Learning.
https://agentos.org
Apache License 2.0
13 stars 4 forks source link

[WIP] PAPAG and SB3 benchmark against the same environment component #370

Closed nickjalbert closed 2 years ago

nickjalbert commented 2 years ago

WIP, some notes:

agentos run agent --entry-point learn --arg-set-file a2c_pong_args.yaml

or CartPole:

agentos run agent --entry-point learn --arg-set-file a2c_cartpole_args.yaml
nickjalbert commented 2 years ago

@andyk FYI, here's the work in progress for SB3 and PAPAG running against the same environment (Pong) component. This should be working now, for PAPAG:

agentos run agent --entry-point learn --arg-set-file ppo_pong_args.yaml
agentos run agent --entry-point evaluate --arg-set-file ppo_pong_args.yaml

and for SB3

agentos run sb3_agent --entry-point learn
agentos run sb3_agent --entry-point evaluate
nickjalbert commented 2 years ago

Closing; new WIP branch is nj_ilya_bench3