add samples

app.py

@@ -7,8 +7,20 @@ import requests
 from tensorflow.keras.models import load_model
 from PIL import Image, ImageDraw
 
-# Load the trained model
-model = load_model('objdet_1_2.h5', custom_objects={'huber_loss': tf.keras.losses.Huber()})
+# Define model path and URL for dynamic downloading
+MODEL_PATH = "objdet_1_2.h5"
+MODEL_URL = "https://huggingface.co/YOUR_USERNAME/objdet_1_2/resolve/main/objdet_1_2.h5"
+
+# Download model if not present
+if not os.path.exists(MODEL_PATH):
+    print("Downloading model...")
+    response = requests.get(MODEL_URL)
+    with open(MODEL_PATH, 'wb') as f:
+        f.write(response.content)
+    print("Model downloaded successfully.")
+
+# Load the model
+model = load_model(MODEL_PATH, custom_objects={'huber_loss': tf.keras.losses.Huber()})
 
 # Constants
 IMAGE_SIZE = (256, 256)
@@ -55,20 +67,42 @@ def draw_bounding_box(image, bbox, label):
 
     return image
 
+# Example images (Replace with actual paths or URLs)
+example_images = [
+    "example1.png",
+    "example2.png",
+    "example3.png",
+    "example4.png"
+]
+
 # Create Gradio interface
-iface = gr.Interface(
-    fn=predict,
-    inputs=gr.Image(type="pil"),
-    outputs=[
-        gr.Text(label="Classification"),
-        gr.Text(label="Confidence Score"),
-        gr.Text(label="Bounding Box (x_min, y_min, x_max, y_max)"),
-        gr.Image(label="Image with Bounding Box")
-    ],
-    title="Seamount Detection",
-    description="Upload an image to classify and detect seamounts. The bounding box is drawn on detected objects."
-)
-
-# Launch the interface
+with gr.Blocks() as demo:
+    gr.Markdown("# Seamount Detection")
+    gr.Markdown("Upload an image or drag one of the example images below to classify and detect seamounts.")
+
+    with gr.Row():
+        input_image = gr.Image(type="pil", label="Upload or Drag an Image")
+        output_image = gr.Image(label="Image with Bounding Box")
+
+    with gr.Row():
+        output_class = gr.Text(label="Classification")
+        output_confidence = gr.Text(label="Confidence Score")
+        output_bbox = gr.Text(label="Bounding Box (x_min, y_min, x_max, y_max)")
+
+    submit_btn = gr.Button("Predict")
+
+    # Example image section
+    gr.Markdown("### Example Images (Drag one into the input box)")
+    with gr.Row():
+        examples = [gr.Image(value=img, type="pil", interactive=True, label=f"Example {i+1}") for i, img in enumerate(example_images)]
+
+    # Connect the prediction function
+    submit_btn.click(predict, inputs=input_image, outputs=[output_class, output_confidence, output_bbox, output_image])
+
+    # Enable dragging example images to input
+    for example in examples:
+        example.change(fn=lambda img: img, inputs=example, outputs=input_image)
+
+# Launch the app
 if __name__ == "__main__":
-    iface.launch()
+    demo.launch()