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
{{ message }}
This repository was archived by the owner on Nov 28, 2025. It is now read-only.
It would be helpful if the user could choose whether to use or not the CUDA acceleration even when CUDA is available in the machine.
For example, in the code I'm running, the model is too big to fit the GPU memory so I have to disable CUDA but the Nemo library has hardcoded checks which enable it and lead to errors.