convert to pil image

app.py

@@ -10,6 +10,8 @@ import PIL
 from PIL import Image
 import numpy as np
 
+import torchvision.transforms.functional as F
+
 output_res = (768,768)
 
 conditioning_image_transforms = T.Compose(
@@ -38,12 +40,12 @@ def infer(prompt, negative_prompt, image):
     # implement your inference function here
     inp = Image.fromarray(image)
 
-    cond_input = conditioning_image_transforms(inp)
+    cond_input = F.to_pil_image(conditioning_image_transforms(inp)[0])
     #cond_input = T.ToPILImage(cond_input)
     
     output = pipe(
         prompt,
-        cond_input[0],
+        cond_input,
         generator=generator,
         num_images_per_prompt=4,
         num_inference_steps=20