[Question] Training Donkey Car Without Simulator Rendering

[Gribesh]

❓ Question

I've been training a reinforcement learning agent for the Donkey Car environment using Stable Baselines3 Zoo. My training command is as follows:

python train.py --algo tqc --env donkey-waveshare-v0 --eval-freq -1 --save-freq 20000 --tensorboard-log ./tblog

I've noticed that the visual rendering of the simulator significantly slows down the training process. I'd like to speed up the training by disabling this simulator rendering. Upon checking the training code, I didn't find any parameter or flag that controls the rendering.

Is there a way to train it without rendering the env? If so, could you please guide me on how to train the agent without the visual overhead of the simulator? Any insights or document recommendations would be highly appreciated!

Thank you

Checklist

• [X] I have checked that there is no similar issue in the repo
• [X] I have read the SB3 documentation
• [X] I have read the RL Zoo documentation
• [X] If code there is, it is minimal and working
• [X] If code there is, it is formatted using the markdown code blocks for both code and stack traces.

[araffin]

Hello,

Is there a way to train it without rendering the env? If so, could you please guide me on how to train the agent without the visual overhead of the simulator?

This is more a question for the donkey car simulation. I tried in the past but without success, however, you can now play the simulation faster than realtime.

[Gribesh]

Oh okay. Thanks for the information.

Actually, I wanted to run the algorithm for a few million steps in the AWS Sagemaker.