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 it possible to do the processing using 16-bit float type, using model.half() and torch.HalfTensor data type? it could lower the VRAM requirements for high resolution images.
The text was updated successfully, but these errors were encountered:
Is it possible to do the processing using 16-bit float type, using model.half() and torch.HalfTensor data type? it could lower the VRAM requirements for high resolution images.
The text was updated successfully, but these errors were encountered: