Update app.py

app.py

@@ -1,77 +1,145 @@
 import gradio as gr
-import torch
-from ultralyticsplus import YOLO, render_result
+import spaces
+from huggingface_hub import hf_hub_download
 
 
-torch.hub.download_url_to_file(
-    'https://external-content.duckduckgo.com/iu/?u=https%3A%2F%2Ftexashafts.com%2Fwp-content%2Fuploads%2F2016%2F04%2Fconstruction-worker.jpg', 'one.jpg')
-torch.hub.download_url_to_file(
-    'https://www.pearsonkoutcherlaw.com/wp-content/uploads/2020/06/Construction-Workers.jpg', 'two.jpg')
-torch.hub.download_url_to_file(
-    'https://nssgroup.com/wp-content/uploads/2019/02/Building-maintenance-blog.jpg', 'three.jpg')
+def download_models(model_id):
+    hf_hub_download("flinta/test", filename=f"{model_id}", local_dir=f"./")
+    return f"./{model_id}"
 
-
-def yoloV8_func(image: gr.inputs.Image = None,
-                image_size: gr.inputs.Slider = 640,
-                conf_threshold: gr.inputs.Slider = 0.4,
-                iou_threshold: gr.inputs.Slider = 0.50):
-    """This function performs YOLOv8 object detection on the given image.
-
-    Args:
-        image (gr.inputs.Image, optional): Input image to detect objects on. Defaults to None.
-        image_size (gr.inputs.Slider, optional): Desired image size for the model. Defaults to 640.
-        conf_threshold (gr.inputs.Slider, optional): Confidence threshold for object detection. Defaults to 0.4.
-        iou_threshold (gr.inputs.Slider, optional): Intersection over Union threshold for object detection. Defaults to 0.50.
+@spaces.GPU
+def yolov9_inference(img_path, model_id, image_size, conf_threshold, iou_threshold):
     """
-    # Load the YOLOv8 model from the 'best.pt' checkpoint
-    model_path = "best.pt"
-    model = YOLO(model_path)
-
-    # Perform object detection on the input image using the YOLOv8 model
-    results = model.predict(image,
-                            conf=conf_threshold,
-                            iou=iou_threshold,
-                            imgsz=image_size)
-
-    # Print the detected objects' information (class, coordinates, and probability)
-    box = results[0].boxes
-    print("Object type:", box.cls)
-    print("Coordinates:", box.xyxy)
-    print("Probability:", box.conf)
-
-    # Render the output image with bounding boxes around detected objects
-    render = render_result(model=model, image=image, result=results[0])
-    return render
-
-
-inputs = [
-    gr.inputs.Image(type="filepath", label="Input Image"),
-    gr.inputs.Slider(minimum=320, maximum=1280, default=640,
-                     step=32, label="Image Size"),
-    gr.inputs.Slider(minimum=0.0, maximum=1.0, default=0.001,
-                     step=0.05, label="Confidence Threshold"),
-    gr.inputs.Slider(minimum=0.0, maximum=1.0, default=0.5,
-                     step=0.05, label="IOU Threshold"),
-]
-
-
-outputs = gr.outputs.Image(type="filepath", label="Output Image")
-
-title = "YOLOv8 101: Custom Object Detection on Construction Workers"
-
-
-examples = [['one.jpg', 640, 0.5, 0.7],
-            ['two.jpg', 800, 0.5, 0.6],
-            ['three.jpg', 900, 0.5, 0.8]]
-
-yolo_app = gr.Interface(
-    fn=yoloV8_func,
-    inputs=inputs,
-    outputs=outputs,
-    title=title,
-    examples=examples,
-    cache_examples=True,
-)
-
-# Launch the Gradio interface in debug mode with queue enabled
-yolo_app.launch(debug=True, enable_queue=True)
+    Load a YOLOv9 model, configure it, perform inference on an image, and optionally adjust 
+    the input size and apply test time augmentation.
+    
+    :param model_path: Path to the YOLOv9 model file.
+    :param conf_threshold: Confidence threshold for NMS.
+    :param iou_threshold: IoU threshold for NMS.
+    :param img_path: Path to the image file.
+    :param size: Optional, input size for inference.
+    :return: A tuple containing the detections (boxes, scores, categories) and the results object for further actions like displaying.
+    """
+    # Import YOLOv9
+    import yolov9
+    
+    # Load the model
+    model_path = download_models(model_id)
+    model = yolov9.load(model_path, device="cuda:0")
+    
+    # Set model parameters
+    model.conf = conf_threshold
+    model.iou = iou_threshold
+    
+    # Perform inference
+    results = model(img_path, size=image_size)
+
+    # Optionally, show detection bounding boxes on image
+    output = results.render()
+    
+    return output[0]
+
+
+def app():
+    with gr.Blocks():
+        with gr.Row():
+            with gr.Column():
+                img_path = gr.Image(type="filepath", label="Image")
+                model_path = gr.Dropdown(
+                    label="Model",
+                    choices=[
+                        "gelan-c.pt",
+                        "gelan-e.pt",
+                        "yolov9-c.pt",
+                        "yolov9-e.pt",
+                    ],
+                    value="gelan-e.pt",
+                )
+                image_size = gr.Slider(
+                    label="Image Size",
+                    minimum=320,
+                    maximum=1280,
+                    step=32,
+                    value=640,
+                )
+                conf_threshold = gr.Slider(
+                    label="Confidence Threshold",
+                    minimum=0.1,
+                    maximum=1.0,
+                    step=0.1,
+                    value=0.4,
+                )
+                iou_threshold = gr.Slider(
+                    label="IoU Threshold",
+                    minimum=0.1,
+                    maximum=1.0,
+                    step=0.1,
+                    value=0.5,
+                )
+                yolov9_infer = gr.Button(value="Inference")
+
+            with gr.Column():
+                output_numpy = gr.Image(type="numpy",label="Output")
+
+        yolov9_infer.click(
+            fn=yolov9_inference,
+            inputs=[
+                img_path,
+                model_path,
+                image_size,
+                conf_threshold,
+                iou_threshold,
+            ],
+            outputs=[output_numpy],
+        )
+        
+        gr.Examples(
+            examples=[
+                [
+                    "data/zidane.jpg",
+                    "gelan-e.pt",
+                    640,
+                    0.4,
+                    0.5,
+                ],
+                [
+                    "data/huggingface.jpg",
+                    "yolov9-c.pt",
+                    640,
+                    0.4,
+                    0.5,
+                ],
+            ],
+            fn=yolov9_inference,
+            inputs=[
+                img_path,
+                model_path,
+                image_size,
+                conf_threshold,
+                iou_threshold,
+            ],
+            outputs=[output_numpy],
+            cache_examples=True,
+        )
+
+
+gradio_app = gr.Blocks()
+with gradio_app:
+    gr.HTML(
+        """
+    <h1 style='text-align: center'>
+    YOLOv9: Learning What You Want to Learn Using Programmable Gradient Information
+    </h1>
+    """)
+    gr.HTML(
+        """
+        <h3 style='text-align: center'>
+        Follow me for more!
+        <a href='https://twitter.com/kadirnar_ai' target='_blank'>Twitter</a> | <a href='https://github.com/kadirnar' target='_blank'>Github</a> | <a href='https://www.linkedin.com/in/kadir-nar/' target='_blank'>Linkedin</a>  | <a href='https://www.huggingface.co/kadirnar/' target='_blank'>HuggingFace</a>
+        </h3>
+        """)
+    with gr.Row():
+        with gr.Column():
+            app()
+
+gradio_app.launch(debug=True)