Update app.py

app.py

@@ -228,22 +228,23 @@ def infer(
             width=width,
             height=height,
          #   latents=None,
-            output_type='latent',
+          #  output_type='latent',
             generator=generator,
             max_sequence_length=512
-        ).images
+        ).images[0]
+        
         print('-- got image --')
-        sd35_image_image = pipe.vae.decode(sd_image / 0.18215).sample
-        sd35_image = sd35_image.cpu().permute(0, 2, 3, 1).float().detach().numpy()
-        sd35_image = (sd35_image * 255).round().astype("uint8")
-        image_pil = Image.fromarray(sd35_image[0])
+        #sd35_image = pipe.vae.decode(sd_image / 0.18215).sample
+       # sd35_image = sd35_image.cpu().permute(0, 2, 3, 1).float().detach().numpy()
+       # sd35_image = (sd35_image * 255).round().astype("uint8")
+       # image_pil = Image.fromarray(sd35_image[0])
+      #  sd35_path = f"sd35_{seed}.png"
+       # image_pil.save(sd35_path,optimize=False,compress_level=0)
+       # upload_to_ftp(sd35_path)
+
         sd35_path = f"sd35_{seed}.png"
-        image_pil.save(sd35_path,optimize=False,compress_level=0)
+        sd_image.save(sd35_path,optimize=False,compress_level=0)
         upload_to_ftp(sd35_path)
-
-    #sd35_path = f"sd35_{seed}.png"
-    #sd_image.save(sd35_path,optimize=False,compress_level=0)
-    #upload_to_ftp(sd35_path)
         # Convert the generated image to a tensor
     #generated_image_tensor = torch.tensor([np.array(sd_image).transpose(2, 0, 1)]).to('cuda') / 255.0
     # Encode the generated image into latents