add negative_scale

gen_img_diffusers.py

@@ -694,8 +694,21 @@ class PipelineLike():
         **kwargs,
     )
 
+    if args.negative_scale is not None:
+      _, real_uncond_embeddings, _ = get_weighted_text_embeddings(
+        pipe=self,
+        prompt=prompt,
+        uncond_prompt=[""]*batch_size,
+        max_embeddings_multiples=max_embeddings_multiples,
+        clip_skip=self.clip_skip,
+        **kwargs,
+    )	
+
     if do_classifier_free_guidance:
+      if args.negative_scale is None:
         text_embeddings = torch.cat([uncond_embeddings, text_embeddings])
+      else:
+      	text_embeddings = torch.cat([uncond_embeddings, text_embeddings, real_uncond_embeddings])
 
     # CLIP guidanceで使用するembeddingsを取得する
     if self.clip_guidance_scale > 0:
@@ -828,16 +841,20 @@ class PipelineLike():
 
     for i, t in enumerate(tqdm(timesteps)):
       # expand the latents if we are doing classifier free guidance
-      latent_model_input = latents.repeat((2, 1, 1, 1)) if do_classifier_free_guidance else latents
+      repeats = 3 if args.negative_scale is not None else 2
+      latent_model_input = latents.repeat((repeats, 1, 1, 1)) if do_classifier_free_guidance else latents
       latent_model_input = self.scheduler.scale_model_input(latent_model_input, t)
-
       # predict the noise residual
       noise_pred = self.unet(latent_model_input, t, encoder_hidden_states=text_embeddings).sample
 
       # perform guidance
       if do_classifier_free_guidance:
+        if args.negative_scale is None:
           noise_pred_uncond, noise_pred_text = noise_pred.chunk(2)
           noise_pred = noise_pred_uncond + guidance_scale * (noise_pred_text - noise_pred_uncond)
+        else:
+          noise_pred_negative, noise_pred_text, noise_pred_uncond = noise_pred.chunk(3)
+          noise_pred = noise_pred_uncond + guidance_scale * (noise_pred_text - noise_pred_uncond) - args.negative_scale	* (noise_pred_negative - noise_pred_uncond)
 
       # perform clip guidance
       if self.clip_guidance_scale > 0 or self.clip_image_guidance_scale > 0 or self.vgg16_guidance_scale > 0:
@@ -2499,6 +2516,8 @@ if __name__ == '__main__':
                       help="1st stage steps for highres fix / highres fixの最初のステージのステップ数")
   parser.add_argument("--highres_fix_save_1st", action='store_true',
                       help="save 1st stage images for highres fix / highres fixの最初のステージの画像を保存する")
+  parser.add_argument("--negative_scale", type=float, default=None,
+                      help="correctly cfg")
 
   args = parser.parse_args()
   main(args)