Closed sideDesert closed 4 months ago
Thanks for the feedback and detailed reproduction explanation! As I explained in #3 , I do see upstream dependency is broken perhaps in Python 3.7 and Python 3.8. I think a more sustainable approach would be we migrate to Python 3.9 or 3.10. I will do some testing after I get a machine with M2 chip. Let's close this issue for now and focus our discussion in #3 and #5 .
I am trying to manually install the packages on the go and run the
thermostat/run_explainer.py
using the bash command provided in the documentation. The first issue I have encountered is that even after trying to install the thermostat library using pip, it throws an errorExpected Behavior
The thermostat library should get installed without throwing any errors
Current Behavior
It throws an error citing that sklearn is deprecated
Possible Solution
Updating to python 3.8 does fix the installation problem for this package, after installing scikit-learn.
Steps to Reproduce
conda create -p ./py37 python=3.7
2.pip install transformers datasets numpy torch tqdm
(I installed them using separate pip commands)pip install thermostat-datasets
Context (Environment)
Pip List
Pip version - 24.0 Conda version - 23.1.0 Env Python Version - 3.7.12