I get different rewards on same env and same steps and same model. Is it normal?

hill-a / stable-baselines

A fork of OpenAI Baselines, implementations of reinforcement learning algorithms

http://stable-baselines.readthedocs.io/

MIT License

4.16k stars 725 forks source link

Closed mmterkc closed 3 years ago

mmterkc commented 3 years ago

I trained rl model but i get different reward values same env and same steps and same model. Is it normal ?

mmterkc commented 3 years ago

my model is ACKTR

Miffyli commented 3 years ago

See the docs on general tips on RL: Results vary wildly between runs. You have to set seeds correctly (see this section).