Update app.py

app.py

@@ -10,25 +10,29 @@ model = torch.jit.load(model_path, map_location=torch.device('cpu'))
 # Define the prediction function
 def predict(image):
     try:
-        print("Predict function called")  # Check if the function is being called
+        print("Predict function called")
+
+        # تغییر اندازه تصویر به اندازه‌ای که مدل نیاز دارد
+        target_size = (3072, 3072)  # یا هر اندازه‌ای که مدل نیاز دارد
+        image = image.resize(target_size)
 
         # Preprocess image
         image = image.convert("RGB")
         input_tensor = torch.from_numpy(np.array(image)).permute(2, 0, 1).unsqueeze(0).float() / 255.0
 
-        print("Image preprocessed")  # Check if preprocessing is successful
+        print("Image preprocessed")
 
         # Run the model
         with torch.no_grad():
             output = model(input_tensor)
         
-        print("Model executed")  # Check if model execution is successful
+        print("Model executed")
 
         # Postprocess the output
         output_image = output.squeeze().permute(1, 2, 0).numpy()
         output_image = (output_image * 255).astype(np.uint8)
         
-        print("Output generated")  # Check if postprocessing is successful
+        print("Output generated")
         return Image.fromarray(output_image)
 
     except Exception as e:
@@ -36,6 +40,7 @@ def predict(image):
         return None
 
 
+
 # Gradio Interface
 iface = gr.Interface(
     fn=predict,