Update app.py

app.py

@@ -99,12 +99,12 @@ class model:
  
  
  nucleus_image = image['image'].convert('L')
- protein_image = image['mask'].split()[3]
+ #protein_image = image['mask'].split()[3]
+ protein_image = image['mask'].convert('L')
 
  to_tensor = T.ToTensor()
  nucleus_image = to_tensor(nucleus_image)
  protein_image = to_tensor(protein_image)
- #protein_image = protein_image
  #stacked_images = torch.stack([nucleus_image, protein_image], dim=0)
  #processed_images = process_image(stacked_images, dataset)
 
@@ -112,10 +112,12 @@ class model:
  #protein_image = processed_images[1].unsqueeze(0)
  #protein_image = protein_image/torch.max(protein_image)
  #protein_image = 1 - protein_image 
-
+ 
  nucleus_image = nucleus_image.unsqueeze(0)
  nucleus_image = process_image(nucleus_image, dataset, 'nucleus')
- protein_image = 1.0*(protein_image > .01)
+ nucleus_image = nucleus_image.unsqueeze(0)
+ protein_image = protein_image.unsqueeze(0).unsqueeze(0)
+ #protein_image = 1.0*(protein_image > .01)
  
 
  formatted_predicted_sequence = run_sequence_prediction(
@@ -168,13 +170,19 @@ with gr.Blocks(theme='gradio/soft') as demo:
  )
 
  with gr.Row().style(equal_height=True):
- nucleus_image = gr.Image(
- source="upload", 
- tool="color-sketch",
- label="Nucleus Image", 
- interactive=True,
- image_mode="RGBA",
- type="pil"
+ #nucleus_image = gr.Image(
+ # source="upload", 
+ # tool="color-sketch",
+ # label="Nucleus Image", 
+ # interactive=True,
+ # image_mode="RGBA",
+ # type="pil"
+ #)
+ nucleus_image = gr.ImageMask(
+ label = "Nucleus Image",
+ interactive = "True",
+ image_mode = "L",
+ brush_color = "#FFFFFF"
  )
  
  with gr.Row():