Upload inference.py with huggingface_hub

inference.py

@@ -3,63 +3,75 @@ import numpy as np
 from PIL import Image
 import os
 
+# ========================
+# Custom layer (مطلوبة)
+# ========================
+from tensorflow.keras.layers import Layer
+
+class CustomScaleLayer(Layer):
+    def __init__(self, scale=1.0, **kwargs):
+        super(CustomScaleLayer, self).__init__(**kwargs)
+        self.scale = scale
+
+    def call(self, inputs):
+        if isinstance(inputs, (list, tuple)):
+            x = tf.add_n(inputs)
+        else:
+            x = inputs
+        return x * self.scale
+
+    def get_config(self):
+        config = super().get_config()
+        config.update({"scale": self.scale})
+        return config
+
+# ========================
+# Load the model
+# ========================
 def load_model(model_path):
-    """
-    Load the pre-trained plant disease classification model
-    """
     try:
-        model = tf.keras.models.load_model(model_path)
+        model = tf.keras.models.load_model(model_path, custom_objects={'CustomScaleLayer': CustomScaleLayer})
         print("Model loaded successfully!")
         return model
     except Exception as e:
         print(f"Error loading model: {e}")
         return None
 
-def preprocess_image(image_path, target_size=(224, 224)):
-    """
-    Preprocess image for model inference
-    """
+# ========================
+# Image preprocessing
+# ========================
+def preprocess_image(image_path, target_size=(299, 299), normalize=True):
     try:
-        # Load image
         img = Image.open(image_path)
-        
-        # Convert to RGB if necessary
         if img.mode != 'RGB':
             img = img.convert('RGB')
-        
-        # Resize image
         img = img.resize(target_size)
-        
-        # Convert to numpy array and normalize
-        img_array = np.array(img) / 255.0
-        
-        # Add batch dimension
+        img_array = np.array(img)
+        if normalize:
+            img_array = img_array / 255.0
         img_array = np.expand_dims(img_array, axis=0)
-        
         return img_array
     except Exception as e:
         print(f"Error preprocessing image: {e}")
         return None
 
+# ========================
+# Prediction
+# ========================
 def predict_disease(model, image_array):
-    """
-    Make prediction on preprocessed image
-    """
     try:
-        # Make prediction
         prediction = model.predict(image_array)
         predicted_class = np.argmax(prediction[0])
         confidence = prediction[0][predicted_class]
-        
         return predicted_class, confidence, prediction[0]
     except Exception as e:
         print(f"Error making prediction: {e}")
         return None, None, None
 
+# ========================
+# Class names
+# ========================
 def get_class_name(class_index):
-    """
-    Get class name from class index
-    """
     classes = [
         "Apple___Apple_scab", "Apple___Black_rot", "Apple___Cedar_apple_rust", "Apple___healthy",
         "Blueberry___healthy", "Cherry_(including_sour)___Powdery_mildew", "Cherry_(including_sour)___healthy",
@@ -75,60 +87,51 @@ def get_class_name(class_index):
         "Tomato___Target_Spot", "Tomato___Tomato_Yellow_Leaf_Curl_Virus", "Tomato___Tomato_mosaic_virus",
         "Tomato___healthy"
     ]
-    
     if 0 <= class_index < len(classes):
         return classes[class_index]
     else:
         return "Unknown"
 
+# ========================
+# Main function
+# ========================
 def main():
-    """
-    Main function to demonstrate model usage
-    """
-    # Model path (update this path to your model location)
     model_path = "Pretrained_model.h5"
-    
-    # Check if model exists
+    sample_image_path = "sample_image.jpg"  # 👈 ضع هنا اسم الصورة
+
     if not os.path.exists(model_path):
         print(f"Model file not found at: {model_path}")
-        print("Please ensure the model file is in the current directory or update the path.")
         return
-    
-    # Load model
+
+    if not os.path.exists(sample_image_path):
+        print(f"Image file not found at: {sample_image_path}")
+        return
+
     model = load_model(model_path)
     if model is None:
         return
-    
-    # Example usage with a sample image
-    # Replace 'sample_image.jpg' with your actual image path
-    sample_image_path = "sample_image.jpg"
-    
-    if os.path.exists(sample_image_path):
-        # Preprocess image
-        image_array = preprocess_image(sample_image_path)
-        if image_array is None:
-            return
-        
-        # Make prediction
-        predicted_class, confidence, all_predictions = predict_disease(model, image_array)
-        
-        if predicted_class is not None:
-            class_name = get_class_name(predicted_class)
-            print(f"\nPrediction Results:")
-            print(f"Predicted Class: {class_name}")
-            print(f"Confidence: {confidence:.2%}")
-            print(f"Class Index: {predicted_class}")
-            
-            # Show top 3 predictions
-            top_3_indices = np.argsort(all_predictions)[-3:][::-1]
-            print(f"\nTop 3 Predictions:")
-            for i, idx in enumerate(top_3_indices):
-                class_name = get_class_name(idx)
-                confidence = all_predictions[idx]
-                print(f"{i+1}. {class_name}: {confidence:.2%}")
-    else:
-        print(f"Sample image not found at: {sample_image_path}")
-        print("Please provide a valid image path to test the model.")
+
+    # تحقق هل الموديل فيه طبقة Rescaling
+    has_rescaling = any(isinstance(layer, tf.keras.layers.Rescaling) for layer in model.layers)
+    image_array = preprocess_image(sample_image_path, target_size=(299, 299), normalize=not has_rescaling)
+    if image_array is None:
+        return
+
+    predicted_class, confidence, all_predictions = predict_disease(model, image_array)
+    if predicted_class is not None:
+        class_name = get_class_name(predicted_class)
+        print(f"\nPrediction Results:")
+        print(f"Predicted Class: {class_name}")
+        print(f"Confidence: {confidence:.2%}")
+        print(f"Class Index: {predicted_class}")
+
+        # Show top 3 predictions
+        top_3_indices = np.argsort(all_predictions)[-3:][::-1]
+        print(f"\nTop 3 Predictions:")
+        for i, idx in enumerate(top_3_indices):
+            class_name = get_class_name(idx)
+            confidence = all_predictions[idx]
+            print(f"{i+1}. {class_name}: {confidence:.2%}")
 
 if __name__ == "__main__":
-    main() 
+    main()