Closed Jetpackjules closed 1 year ago
Thanks!
But what Rlib? is it similar?
I do not know about Rllib as I am unable to use it on my pc, sorry.
For rllib you should be able to launch with the same command line args used for training and add --eval
For rllib you should be able to launch with the same command line args used for training and add
--eval
Does this still work with just gdrl
?
Also, is there a way to pick a previously trained model? (The ones that have their logs saved in the ray_wrapper folder...)
Hey @Jetpackjules , did you resolve this?
Not exactly, I just switched to Sb3.
Also, I did not have time to reply before the thread was closed, but I read the hugging face documentation, which helped a LOT with my project! Thank you for this awesome plugin!
Ok well, that is good I guess. I will close this issue then.
You can check https://github.com/ryash072007/godot_rl_agents_forked/blob/main/godot_rl/backend/sb3.py.