init

app.py

@@ -19,7 +19,7 @@ def infer_gpu_part(pipe, seed, prompt, negative_prompt, ddim_steps, guidance_sca
     result = pipe(prompt, negative_prompt=negative_prompt, generator=generator,
                 num_inference_steps=ddim_steps, guidance_scale=guidance_scale,
                 resolutions_list=resolutions_list, fast_mode=fast_mode, cosine_scale=cosine_scale,
-                )
+                ).images[0]
     return result
 
 def infer(prompt, output_size, ddim_steps, guidance_scale, cosine_scale, seed, options, negative_prompt):
@@ -43,9 +43,8 @@ def infer(prompt, output_size, ddim_steps, guidance_scale, cosine_scale, seed, o
     result = infer_gpu_part(pipe, seed, prompt, negative_prompt, ddim_steps, guidance_scale, resolutions_list, fast_mode, cosine_scale, disable_freeu)
     print('GPU ends')
 
-    image = result.images[0]
     save_path = 'output.png'
-    image.save(save_path)
+    result.save(save_path)
 
     return save_path
 

pipeline_freescale.py

@@ -1131,6 +1131,8 @@ class StableDiffusionXLPipeline(DiffusionPipeline, FromSingleFileMixin, LoraLoad
         if not return_dict:
             return (image,)
 
+        print(image.shape)
+
         return StableDiffusionXLPipelineOutput(images=image)
 
     # Overrride to properly handle the loading and unloading of the additional text encoder.