Try to keep text encoders loaded and patched to increase speed.
load_model_gpu() is now used with the text encoder models instead of just the unet.
Showing
Please register or sign in to comment
load_model_gpu() is now used with the text encoder models instead of just the unet.