Update app.py

app.py

@@ -24,6 +24,9 @@ with gr.Blocks() as demo:
     def modelTraining(img):
         class_names = ['T-shirt/top', 'Trouser', 'Pullover', 'Dress', 'Coat',
                'Sandal', 'Shirt', 'Sneaker', 'Bag', 'Ankle boot']
+    
+        # Normalize the pixel values
+        img = np.array(img) / 255.0
         
         # clothing dataset
         mnist = tf.keras.datasets.mnist
@@ -66,14 +69,7 @@ with gr.Blocks() as demo:
         # create the final model for production
         probability_model = tf.keras.Sequential([model, tf.keras.layers.Softmax()])
 
-        # Input image pre-processing before submission to the model
-        # the image can be passed as a PIL or numpy
-
-        # Normalize the pixel values?
-        print(f"Input image shape: {img.shape}  Dimensions: {img.ndim} Array Element: {img[0]} ***********************************************************************")
-        # assuming image_array is your input image array of shape (552, 3)
-        resized_array = np.resize(img, (28, 28)) # resize the array to (28, 28)
-        input_array = np.expand_dims(resized_array, axis=0) # add an extra dimension to represent the batch size
+        input_array = np.expand_dims(img, axis=0) # add an extra dimension to represent the batch size
 
         # Make a prediction using the model
         prediction = probability_model.predict(input_array)
@@ -85,7 +81,7 @@ with gr.Blocks() as demo:
     # Creates the Gradio interface objects
     with gr.Row():
         with gr.Column(scale=2):
-            image_data = gr.Image(label="Upload Image", type="numpy")
+            image_data = gr.Image(label="Upload Image", type="numpy", image_mode="L", shape=[28,28], invert_colors=True)
         with gr.Column(scale=1):
             model_prediction = gr.Text(label="Model Prediction", interactive=False)
     image_data.change(modelTraining, image_data, model_prediction)