Train with lower precision
mhusseinsh opened this issue · comments
Hello,
I would like to ask if there is a way to train the UNIT model with a lower precision? just to to give it a try if I can result in reducing training/inference time
Which part do I need to edit in the code?
The updated UNIT implementation is available at
https://github.com/NVlabs/imaginaire
We have low precision training.
Support should be better there.