Toby Kim
Toby Kim
@tamarott Does this mean that we can't used a SR model that is trained on imageA to super-resolve for imageB, C, D...etc.?
works for conda env with python 3.7, tf-gpu 2.3.0, all other modules were installed without version specifications using pip.
TLDR: you need to use pallas splash attention kernel to save memory and boost speeds on TPUs (no other method worked for me) i've looked into this topic heavily (on...
Awesome :) Qwen 3-4B and 8B would be my primary interest. I'm planning to use as a backbone for a speech model, through TPUs provided by the TRC. Bigger models...