You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Wanted to bring to your notice that the installation from the environment.yml file seems to install cpu versions of pytorch and torchvision instead of gpu ones. I found that this is because it installs pytorch and torchvision from the conda-forge channel instead of the pytorch channel. I managed to resolve this issue by installing pytorch and torchvision separately from the pytorch channel.
I was also able to reproduce this issue on the docker installation since that makes use of the same yml file.
The text was updated successfully, but these errors were encountered:
Thanks for reporting this! I think the available packages changed at some point which led to the conda-forge version getting pulled instead of the pytorch version. We'll update the environment file to make sure the correct package is installed.
Hi,
Thanks for the great framework!
Wanted to bring to your notice that the installation from the
environment.yml
file seems to installcpu
versions ofpytorch
andtorchvision
instead ofgpu
ones. I found that this is because it installspytorch
andtorchvision
from theconda-forge
channel instead of thepytorch
channel. I managed to resolve this issue by installingpytorch
andtorchvision
separately from thepytorch
channel.I was also able to reproduce this issue on the docker installation since that makes use of the same yml file.
The text was updated successfully, but these errors were encountered: