Ilya Kostrikov

Try to use a stable release. They are going to introduce a lot of changes in the more recent version (the one that is currently developed in the master branch).

0.3

I'm using the architecture from open ai universe starter. It works similarly. But it was easier to use it because they provide all hyper parameters.

Check the instructions in pytorch.org. There is 0.3 there.

@dragen1860 it's designed specifically to be efficient on CPU. For a GPU optimized code see http://github.com/ikostrikov/pytorch-a2c-ppo-acktr

Looks good to me! Could you please squash commits into one (or 2-3) and use a message like "Add tensorboard logging" and I will accept a PL?

Thanks! Are you planning to change anything or I can merge the PL?

Hi, sorry for delays (was busy with the NIPS deadline). I will review asap.

Blocking doesn't happen to me. What configuration are you using?

Please report more information. I tested it on ubuntu 16.04. PyTorch 0.2 and 0.3, python 3.6 and it works for me both on ubuntu and os x.