search

DLR-RM / rl-baselines3-zoo

A training framework for Stable Baselines3 reinforcement learning agents, with hyperparameter optimization and pre-trained agents included.
https://rl-baselines3-zoo.readthedocs.io
MIT License
1.9k stars 495 forks source link

[Feature Request] Specify custom keyword arguments for eval environments #404

Closed Quentin18 closed 9 months ago

Quentin18 commented 9 months ago

🚀 Feature

Add the --eval-env-kwargs parameter to the train.py script in order to specify keyword arguments to pass to the env constructor for evaluation.

Motivation

Actually, we cannot specify different keyword arguments between training and evaluation environments. It can be usefull if we want the evaluation environment to behave differently from the training one.

Pitch

No response

Alternatives

No response

Additional context

You can see the usage in this commit.

Checklist

araffin commented 9 months ago

Hello, Sounds like a good to have feature, i would be happy to receive a PR for it =) (the default should remain the same though)

Quentin18 commented 9 months ago

Hello, OK thanks :ok_hand: ! Here is the PR : https://github.com/DLR-RM/rl-baselines3-zoo/pull/405