search

shariqiqbal2810 / maddpg-pytorch

PyTorch Implementation of MADDPG (Lowe et. al. 2017)
MIT License
569 stars 129 forks source link

Question about some arguments. #3

Closed ShawnLue closed 5 years ago

ShawnLue commented 5 years ago

Hi Shariq,

I'm a little confused about the 'n_rollout_threads' and 'n_training_threads'.

In your code, n_rollout_threads is set to 1. If I understand correctly, this indicates using DummyVecEnv without collecting experiences in parallel. Any insight why not using SubprocVecEnv like in a2c?

And what about n_training_threads? Is this setting only for acceleration?

Thanks a lot!

shariqiqbal2810 commented 5 years ago

Hi Shawn,

n_rollout_threads is set to 1 by default, but you can set it to however many available threads you have on your machine. For example, I used 12.

n_training_threads just controls how many threads PyTorch uses for inference if you're not using a GPU.

ShawnLue commented 5 years ago

Thank you so much!