I am facing below issue:
I updated pytorch version to 0.2.0 still its showing in code its using pytorch version 0.1.12
I abserved in conda list pytorch-gpu version 0.1.12
pytorch version is 0.2.0
Is it using pytorch-gpu?
I updated pytorch version to 0.2.0 still its showing in code its using pytorch version 0.1.12
How can i fix that issue ?
How did you install pytorch? If it was through conda, what does your conda list say?
I was installed using conda install -c soumith pytorch
here my conda list:
In that list pytorch version is 0.2.0 but in my code when I checked with torch.version its giving me 0.1.12
1 Like
What is pytorch-gpu? Doesn’t seem like you should have it/need it since pytorch itself has gpu capabilities. I’d recommend a conda uninstall pytorch-gpu and seeing where that gets you.
What does your torch.version say now?
When I am trying to find torch.__version __ its giving me AttributeError: module torch has no attribute ‘version’ and above ImportError .
Do I need to install pytorch from source
And what happens if you conda uninstall pytorch and then conda install pytorch?
I’m not really sure what the problem is but I’d recommend reinstalling pytorch as above (it doesn’t have to be from source).
I was re installed pytorch as you said and unistalled pytorch-gpu then when I am checking torch version in code its giving me above error for version module.
con’t import name variable from torch.autograd