Update app.py

app.py

@@ -1,44 +1,48 @@
-import gradio as gr
-import requests
-from tensorflow import keras
-
-def classify_image(input_image):
-    # Download human-readable labels for ImageNet.
-    try:
-        response = requests.get("https://git.io/JJkYN")
-        response.raise_for_status()  # Ensure the request was successful
-        labels = response.text.split("\n")
-    except Exception as e:
-        print("Error fetching labels:", e)
-        labels = ["Unknown"] * 1000  # Fallback in case the request fails
-
-    # Load the MobileNetV2 model
-    inception_net = keras.applications.MobileNetV2(
-        input_shape=(224, 224, 3),
-        alpha=1.0,
-        include_top=True,
-        weights="imagenet",
-        classes=1000,
-        classifier_activation="softmax"
-    )
-
-    # Resize the input image to the expected size
-    input_image = input_image.reshape((1, 224, 224, 3))  # Reshape for a single prediction
-    input_image = keras.applications.mobilenet_v2.preprocess_input(input_image)
-    
-    # Perform prediction
-    prediction = inception_net.predict(input_image).flatten()
-    confidences = {labels[i]: float(prediction[i]) for i in range(len(labels))}
-    return confidences
-
-demo = gr.Interface(
-    fn=classify_image,
-    inputs=gr.Image(interactive=True, label="Upload Image"), 
-    outputs=gr.Label(num_top_classes=3, label="Top Predictions"),
-    examples=["./images/banana.jpg", "./images/car.jpg", "./images/guitar.jpg", "./images/lion.jpg"],  # Use valid URLs or local paths
-    theme="default",
-    css=".footer{display:none !important}"
-)
-
-if __name__ == "__main__":
+import requests
+import tensorflow as tf
+import gradio as gr
+
+def classify_image(input_image):
+    # Download human-readable labels for ImageNet.
+    try:
+        response = requests.get("https://git.io/JJkYN")
+        response.raise_for_status()  # Ensure the request was successful
+        labels = response.text.split("\n")
+    except Exception as e:
+        print("Error fetching labels:", e)
+        labels = ["Unknown"] * 1000  # Fallback in case the request fails
+
+    # Load the MobileNetV2 model
+    inception_net = keras.applications.MobileNetV2(
+        input_shape=(224, 224, 3),
+        alpha=1.0,
+        include_top=True,
+        weights="imagenet",
+        classes=1000,
+        classifier_activation="softmax"
+    )
+
+    # Resize the input image to the expected size
+    input_image = input_image.reshape((1, 224, 224, 3))  # Reshape for a single prediction
+    input_image = keras.applications.mobilenet_v2.preprocess_input(input_image)
+    
+    # Perform prediction
+    prediction = inception_net.predict(input_image).flatten()
+    confidences = {labels[i]: float(prediction[i]) for i in range(len(labels))}
+    return confidences
+    
+image = gr.Image(interactive=True, label="Upload Image")
+label = gr.Label(num_top_classes=3, label="Top Predictions")
+
+demo = gr.Interface(
+    title="Image Classifier Keras",
+    fn= classify_image,
+    inputs= image, 
+    outputs= label,
+    examples= [["./images/banana.jpg"], ["./images/car.jpg"], ["./images/guitar.jpg"], ["./images/lion.jpg"]],  # Use valid URLs or local paths
+    theme= "default",
+    css= ".footer{display:none !important}"
+)
+
+if __name__ == "__main__":
     demo.launch(share=True)