Update app.py

app.py

@@ -3,28 +3,36 @@ import tensorflow as tf
 import numpy as np
 from PIL import Image
 
-# Load your model
+# Load the model (ensure you have the correct model path)
 model = tf.keras.models.load_model("denis_mnist_cnn_model.h5")
 
-# Define the prediction function
+# Define a function to preprocess input and make predictions
 def predict(image):
-    image = np.array(image)  # Convert to numpy array
-    image = tf.image.resize(image, (224, 224))  # Resize to the model's expected input size
-    image = np.expand_dims(image, axis=0)  # Add batch dimension (model expects a batch of images)
-    image = image / 255.0  # Normalize pixel values
+    # Convert image to a numpy array
+    image = np.array(image)
     
-    # Check if the model needs flattening
-    if len(image.shape) == 4:  # Check if image has a batch dimension
-        image = tf.keras.layers.Flatten()(image)  # Flatten the image if necessary
+    # Resize the image to the expected shape (28, 28, 3)
+    image = tf.image.resize(image, (28, 28))  # Resize to 28x28 pixels
+    image = np.expand_dims(image, axis=-1)  # Add the channel dimension if grayscale
+    image = np.repeat(image, 3, axis=-1)  # Convert grayscale to RGB (if model was trained on RGB images)
     
+    # Normalize the image
+    image = image / 255.0
+    
+    # Add batch dimension
+    image = np.expand_dims(image, axis=0)  # Add batch dimension to match the model's expected input shape (1, 28, 28, 3)
+    
+    # Perform prediction
     prediction = model.predict(image)
+    
+    # Return prediction as JSON
     return {"prediction": prediction.tolist()}
 
-# Create the Gradio interface
+# Create a Gradio interface
 interface = gr.Interface(
-    fn=predict,
-    inputs="image",  # Image input
-    outputs="json",  # Output as JSON
+    fn=predict, 
+    inputs="image",  # Image input for testing
+    outputs="json"  # JSON output for prediction results
 )
 
 # Launch the interface