HumanCompatibleAI / evaluating-rewards

Library to compare and evaluate reward functions
https://arxiv.org/abs/2006.13900
Apache License 2.0
61 stars 7 forks source link

Conflicting versions of stable-baselines and imitation #55

Open dashpritam opened 1 year ago

dashpritam commented 1 year ago

How do I resolve the following conflict? Please help.

I have tried both imitation@tf-master and tf-master-compatibility branch

Linux, Python 3.7 Tensorflow 1.15

The conflict is caused by:
    evaluating-rewards 0.1.1 depends on stable-baselines 2.10.3a0.-WIP- (from git+https://github.com/hill-a/stable-baselines.git)
    imitation 0.1.1 depends on stable-baselines~=2.10.1

To fix this you could try to:
1. loosen the range of package versions you've specified
2. remove package versions to allow pip attempt to solve the dependency conflict
kierad commented 12 months ago

Hello, I'm also seeing this error when installing this repo.

> python -V
Python 3.7.17
> pip list
Package    Version
---------- -------
pip        23.1.2
setuptools 47.1.0

OS: MacOS Ventura 13.4.1

sibiraja commented 8 months ago

I also am getting the same issue. I attempted to solve the problem by creating forks of both the evaluating_rewards repo and and stable_baselines repo, and changing the version string in stable-baselines/stable_baselines/version/txt, and setting the dependency in evaluating-rewards/requirements.txt to point to my forked version of stable-baselines. This seems to fix the initial issue that you pointed out. However, I need to downgrade my python version to 3.7 in order to install tensorflow 1.15, but homebrew does not support python 3.7. Not sure if there is a workaround beyond a manual installation of python 3.7

sibiraja commented 8 months ago

Update, the best workaround seems to be using a Docker container as mentioned by Dr. Gleave here: