guided-diffusion
guided-diffusion copied to clipboard
If I want to dynamically adjust the learning rate, how should I use the parameter lr_anneal_steps? And how should I set the initial value?
