search

google-research / batch_rl

Offline Reinforcement Learning (aka Batch Reinforcement Learning) on Atari 2600 games
https://offline-rl.github.io/
Apache License 2.0
536 stars 75 forks source link

Python version? #21

Closed johmicrot closed 3 years ago

johmicrot commented 3 years ago

Can we get the python version used in the documentation? I'm trying to implement this in Ubuntu 20 so I need to manually install a previous version of python.

Would also be helpful to the the versions used for absl-py, atari-py, gin-config, opencv-python, gym, numpy. As well as thewhich virtual environment manager that was used (e.g. conda, venv)

agarwl commented 3 years ago

I don't remember which python 3 version was this but python3.6 or python3.7 would probably work -- this code was released in 2019, so I think some of the newer python versions didn't exist back then and python2 wasn't outdated (yet).

Can you possibly send a pull request regarding the specific versions required to run this code -- my guess is that you can use the versions which were available in 2019 (what specifically breaks with the newer versions?)

glorgao commented 3 years ago

python 3.7.9 is ok.

I install the absl-py, atari-py ... by requirements in dopamine-rl==3.1.13, and they all work.