correct example

examples/research_projects/lora/train_text_to_image_lora.py

@@ -479,7 +479,6 @@ def main():
         weight_dtype = torch.bfloat16
 
     if args.use_peft:
-
         from peft import LoraConfig, LoraModel, get_peft_model_state_dict, set_peft_model_state_dict
 
         UNET_TARGET_MODULES = ["to_q", "to_v", "query", "value"]
@@ -496,7 +495,6 @@ def main():
 
         vae.requires_grad_(False)
         if args.train_text_encoder:
-
             config = LoraConfig(
                 r=args.lora_text_encoder_r,
                 lora_alpha=args.lora_text_encoder_alpha,
@@ -506,7 +504,6 @@ def main():
             )
             text_encoder = LoraModel(config, text_encoder)
     else:
-
         # freeze parameters of models to save more memory
         unet.requires_grad_(False)
         vae.requires_grad_(False)