app.py

import gradio as gr
import spaces
from huggingface_hub import hf_hub_download


def download_models(model_id):
    hf_hub_download("merve/yolov9", filename=f"{model_id}", local_dir=f"./")
    return f"./{model_id}"

@spaces.GPU
def yolov9_inference(img_path, model_id, image_size, conf_threshold, iou_threshold):
	"""
	Performs object detection using a YOLOv9 model. This function loads a specified YOLOv9 model,
	configures it based on the provided parameters, and carries out inference on a given image. 
	Additionally, it allows for optional modification of the input size and the application of 
	test time augmentation to potentially improve detection accuracy.

	Parameters:
	    model_path (str): The file path to the pre-trained YOLOv9 model.
	    conf_threshold (float): The confidence threshold used during Non-Maximum Suppression (NMS)
		to filter detections. Detections with confidence levels below this threshold are discarded.
	    iou_threshold (float): The Intersection over Union (IoU) threshold applied in NMS. Detections 
		with IoU values above this threshold are considered overlapping and are hence merged.
	    img_path (str): The file path to the image on which inference is to be performed.
	    size (int, optional): The input size for inference. Altering this may affect the accuracy and 
		speed of the detection process.

	Returns:
	    tuple: A tuple containing the detection results. This includes the bounding boxes (boxes),
	    confidence scores (scores), and detected categories (categories), alongside a results object
	    for further processing, such as visualization.
	"""
	
    # 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=[
                [
                    "image_data/IMG_3352.JPG",
                    "gelan-e.pt",
                    640,
                    0.4,
                    0.5,
                ],
                [
                    "image_data/IMG_3353.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,
        )

# Initialize a Gradio Blocks application.
gradio_app = gr.Blocks()

with gradio_app:
    # Display a title using HTML, centered.
    gr.HTML("""
        <h1 style='text-align: center; margin-bottom: 20px;'>
        YOLOv9 from PipYoloV9 on my data
        </h1>
    """)

    # Assuming 'app()' represents your main app functionality.
    # If 'app()' involves Gradio interface components, directly place those components here.
    # Example: gr.Image() for image input, gr.Button() for a button, etc.
    # Replace 'app()' with your actual Gradio components or function call.
    app()

# Launch the Gradio app, enabling debug mode for detailed error logs and server information.
gradio_app.launch(debug=True)