Open crowsonkb opened 3 years ago
They don't seem to be documented either in this repo or in improved-diffusion. I'd like to match the training flags as closely as possible when fine-tuning (I'm trying to fine-tune the 512x512 ImageNet model).
They don't seem to be documented either in this repo or in improved-diffusion. I'd like to match the training flags as closely as possible when fine-tuning (I'm trying to fine-tune the 512x512 ImageNet model).