liuzuxin/fsrl

🚀 A fast safe reinforcement learning library in PyTorch

fsrl's Issues

关于自定义环境

Required prerequisites

• I have read the documentation https://safety-gymnasium.readthedocs.io.
• I have searched the Issue Tracker and Discussions that this hasn't already been reported. (+1 or comment there if it has.)
• Consider asking first in a Discussion.

Questions

请问我如何使用自定义的环境？

[BUG] "Render" parameter in BaseAgent has unclear meaning.

Required prerequisites

• I have read the documentation https://fsrl.readthedocs.io.
• I have searched the Issue Tracker and Discussions that this hasn't already been reported. (+1 or comment there if it has.)
• Consider asking first in a Discussion.

What version of FSRL are you using?

0.1.0

System and environment information

I guess this is not relevant for this specific problem, but I installed fsrl according to the instructions here on github and I'm using python version 3.8

Problem description

In BaseAgent "collect" method the render parameter is described as "render (bool) – Whether to render the environment during evaluation, defaults to False". However looking at the code, the render is later called only in the collector's (FastCollector in this case) "collect" method, but there the meaning of render is "render (float) – the sleep time between rendering consecutive frames. Default to None (no rendering)."

Reproducible example code

This can be observed for example when running eval_ppol_agent.py from examples folder.

Traceback

No response

Expected behavior

No response

Additional context

No response