Create app.py

app.py

@@ -0,0 +1,53 @@
+import gradio as gr
+import tensorflow as tf
+from tensorflow.keras.applications import EfficientNetV2L
+from tensorflow.keras.applications.efficientnet_v2 import preprocess_input, decode_predictions
+import numpy as np
+from PIL import Image
+
+# Lazy loading to optimize memory usage
+model = None
+
+def load_model():
+    """Load the EfficientNetV2L model only when needed."""
+    global model
+    if model is None:
+        model = EfficientNetV2L(weights="imagenet")
+
+def preprocess_image(image):
+    """Preprocess the image for EfficientNetV2L model inference."""
+    image = image.resize((480, 480))  # Resize for EfficientNetV2L
+    image_array = np.array(image)  # Convert to NumPy array
+    image_array = preprocess_input(image_array)  # Normalize input
+    image_array = np.expand_dims(image_array, axis=0)  # Add batch dimension
+    return image_array
+
+def predict_image(image):
+    """
+    Process the uploaded image and return the top 3 predictions.
+    """
+    try:
+        load_model()  # Ensure the model is loaded
+        image_array = preprocess_image(image)  # Preprocess image
+        predictions = model.predict(image_array)  # Get predictions
+        decoded_predictions = decode_predictions(predictions, top=3)[0]
+
+        # Format predictions as a dictionary (label -> confidence)
+        return {label: float(confidence) for _, label, confidence in decoded_predictions}
+
+    except Exception as e:
+        return {"Error": str(e)}
+
+# Create the Gradio interface
+interface = gr.Interface(
+    fn=predict_image,
+    inputs=gr.Image(type="pil"),  # Accepts an image input
+    outputs=gr.Label(num_top_classes=3),  # Shows top 3 predictions
+    title="EfficientNetV2L Image Classifier",
+    description="Upload an image, and the model will predict its content with high accuracy.",
+    allow_flagging="never"  # Disable flagging to avoid unnecessary logs
+)
+
+# Launch the Gradio app
+if __name__ == "__main__":
+    interface.launch()