Update app.py

app.py

@@ -8,20 +8,39 @@ import cv2
 
 class RobustDepthTo3DConverter:
     def __init__(self):
-        # Load MiDaS model with explicit configuration
+        # Load MiDaS model with explicit configuration and error handling
         self.device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
         
-        # Use a specific version and add error handling
         try:
-            self.model = torch.hub.load("intel-isl/MiDaS", "DPT_Small", pretrained=True)
+            # Try multiple methods to load the model
+            try:
+                # Method 1: Using torch hub with explicit version
+                self.model = torch.hub.load("intel-isl/MiDaS", "MiDaS_small", pretrained=True)
+            except Exception as e1:
+                try:
+                    # Method 2: Alternative loading
+                    self.model = torch.hub.load("intel-isl/MiDaS", "DPT_Small", pretrained=True)
+                except Exception as e2:
+                    # If both methods fail, raise a comprehensive error
+                    raise ValueError(f"Model loading failed. Errors: {e1}, {e2}")
+            
+            # Move model to device and set to eval mode
             self.model.to(self.device)
             self.model.eval()
+            
+            # Verify model is loaded correctly
+            if self.model is None:
+                raise ValueError("Model initialization failed: model is None")
+            
+            print("MiDaS model successfully initialized")
+        
         except Exception as e:
-            print(f"Model loading error: {e}")
+            print(f"Critical model initialization error: {e}")
             self.model = None
         
         # Create transformation pipeline
         self.transform = transforms.Compose([
+            transforms.Resize((256, 256)),  # Standardize input size
             transforms.ToTensor(),
             transforms.Normalize(
                 mean=[0.485, 0.456, 0.406], 
@@ -37,20 +56,6 @@ class RobustDepthTo3DConverter:
         if not isinstance(input_image, Image.Image):
             input_image = Image.fromarray(input_image)
         
-        # Resize image with aspect ratio preservation
-        max_size = 800
-        width, height = input_image.size
-        
-        # Calculate new dimensions while preserving aspect ratio
-        if width > height:
-            new_width = max_size
-            new_height = int(height * (max_size / width))
-        else:
-            new_height = max_size
-            new_width = int(width * (max_size / height))
-        
-        input_image = input_image.resize((new_width, new_height), Image.LANCZOS)
-        
         # Convert to RGB if needed
         if input_image.mode != 'RGB':
             input_image = input_image.convert('RGB')
@@ -62,7 +67,7 @@ class RobustDepthTo3DConverter:
         More robust depth estimation
         """
         if self.model is None:
-            raise ValueError("MiDaS model not properly initialized")
+            raise ValueError("MiDaS model not properly initialized. Check model loading.")
         
         try:
             # Preprocess image
@@ -96,7 +101,7 @@ class RobustDepthTo3DConverter:
     
     def create_point_cloud(self, image, depth_map):
         """
-        Create point cloud with enhanced error handling and adaptive sampling
+        Create point cloud with enhanced error handling
         """
         if depth_map is None:
             return None
@@ -170,6 +175,10 @@ class RobustDepthTo3DConverter:
         """
         Enhanced full pipeline with comprehensive error handling
         """
+        # First, check if model is initialized
+        if self.model is None:
+            raise ValueError("MiDaS model initialization failed. Cannot process image.")
+        
         try:
             # Preprocess and validate input
             input_image = self.preprocess_image(input_image)
@@ -198,7 +207,7 @@ class RobustDepthTo3DConverter:
         
         except Exception as e:
             print(f"Full pipeline error: {e}")
-            return f"Error during conversion: {str(e)}"
+            raise  # Re-raise the exception to be caught in the Gradio interface
 
 def create_huggingface_space():
     # Initialize converter
@@ -206,13 +215,15 @@ def create_huggingface_space():
     
     def convert_image(input_image):
         try:
+            # Check model initialization before processing
+            if converter.model is None:
+                raise ValueError("MiDaS model failed to initialize. Cannot process image.")
+            
             output_model = converter.process_image(input_image)
-            if output_model.startswith("Error"):
-                raise ValueError(output_model)
             return output_model
         except Exception as e:
             print(f"Conversion error: {e}")
-            return str(e)
+            raise gr.Error(f"Conversion failed: {str(e)}")
     
     # Gradio Interface
     iface = gr.Interface(