score_sde_pytorch icon indicating copy to clipboard operation
score_sde_pytorch copied to clipboard

Python 3.7 vs 3.8

Open KyleM4t1qbit opened this issue 2 years ago • 4 comments

Not an issue (I suppose it could be an issue with the documentation), but putting this here in case it helps anyone else:

Running this with Python 3.7 resulted in a segmentation fault during the JIT compilation of the Pytorch C++ extensions.

Using Python 3.8 works without issue (using all packages in requirements.txt).

Hope this helps someone else.

KyleM4t1qbit avatar Apr 20 '22 18:04 KyleM4t1qbit

Thank you very much! It is quite kind of you! I try almost all the versions for Pytorch and none of them works. Finally, the problems are solved after replacing python 3.7 with python 3.8.

mo666666 avatar Feb 16 '23 14:02 mo666666

Thank you very much! It is quite kind of you! I try almost all the versions for Pytorch and none of them works. Finally, the problems are solved after replacing python 3.7 with python 3.8.

Hi, I am struggling with the environment setup would you let me know what is your environment package version like the torch, tensorflow-probability and how you install it as I followed the pip install -r requirment.txt it always came with error. Thanks.

KeepNoob avatar Nov 16 '23 18:11 KeepNoob

Hello, I'm sorry to bother you. I am currently in Python 3.8.18 , but I still encounter import errors. May I ask which version of Python you are referring to. ImportError: cannot import name 'ParamSpec' from 'typing_extensions'

1808030112 avatar Dec 22 '23 08:12 1808030112

I have the same problem. I have tried different python versions.

mengchuangji avatar Dec 28 '23 16:12 mengchuangji