Is it safe to ignore "WARNING T5 model is using fp8 weights for caching. This may affect the quality of the cached outputs."?
#17
by
oddball516
- opened
I'm training flux1-dev lora with fp8 model, I received this with sd-scripts:
WARNING T5 model is using fp8 weights for caching. This may affect the quality of the cached outputs.
Is it safe to ignore it?