Yicheng Luo

I think, in this case, you can consider creating a special agent for your problem that only calls `learner.step` at the end of an episode. Alternatively, there is always the...

@alexis-jacq FYI since we had a discussion last time about some implementation differences between this TD3 and the official PyTorch implementation.

@rdevon The environment_factory should create a new environment every time it is called. Therefore, the examples `rl_discrete/` are indeed all incorrect.

Not sure what the tests failure are about tho.

> We use specific versions of packages to make sure new releases of dependencies don't break things at a random (which happens a number of times). Could you use specific...

@rdevon agree that having some examples would be nice. Fortunately it's not that difficult to do once you get a hang of it. The code for loading back the learner's...

I think the issue here is that the default D4PG networks are unable to handle the observations returned by the NLE environment. The observation space for the NLE environment (obtained...

It's interesting that this didn't work out of the box. I suspect the issue is with buffered writes. Can you try closing the logger when you finish writing to see...

@rdevon The flushing should be done automatically by TB but maybe that's not working in your case. There is a auto-close wrapper https://github.com/deepmind/acme/blob/master/acme/utils/loggers/auto_close.py that you can wrap the logger which...

I think it may be possible to refactor the core_dependencies https://github.com/deepmind/acme/blob/master/setup.py#L49 in the setup.py to include the minimal, allowing you to install acme as a standalone package where you can...