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
Is your feature request related to a problem? Please describe.
I have to install python 3.7 if I want to guarantee compatibility with neox libraries. 3.8 if I feel brave. 3.10 is right out. 3.7 is deprecated later this year.
Describe the solution you'd like
All dependencies are updated to run correctly python 3.10 and current torch and cuda.
Describe alternatives you've considered
Maintaining legacy 3.7 support after it's officially in end of life, backporting new code that doesn't compile on older pytorch versions into the older pytorch versions that neox assumes.
Additional context
I already have a draft pr for this here. I figured I should document it an issue since that's established practice in contributing.md now. Branch is waiting on more testing and the ability to verify that there are no important regressions.
The text was updated successfully, but these errors were encountered:
Is your feature request related to a problem? Please describe.
I have to install python 3.7 if I want to guarantee compatibility with neox libraries. 3.8 if I feel brave. 3.10 is right out. 3.7 is deprecated later this year.
Describe the solution you'd like
All dependencies are updated to run correctly python 3.10 and current torch and cuda.
Describe alternatives you've considered
Maintaining legacy 3.7 support after it's officially in end of life, backporting new code that doesn't compile on older pytorch versions into the older pytorch versions that neox assumes.
Additional context
I already have a draft pr for this here. I figured I should document it an issue since that's established practice in contributing.md now. Branch is waiting on more testing and the ability to verify that there are no important regressions.
The text was updated successfully, but these errors were encountered: