Detomo/Object_detection

1import gradio as gr2import torch3from sahi.prediction import ObjectPrediction4from sahi.utils.cv import visualize_object_predictions, read_image5from ultralyticsplus import YOLO, render_result6 7 8def yolov8_inference(9    image,10    model_path,11    image_size,12    conf_threshold,13    iou_threshold,14):15    """16    YOLOv8 inference function17    Args:18        image: Input image19        model_path: Path to the model20        image_size: Image size21        conf_threshold: Confidence threshold22        iou_threshold: IOU threshold23    Returns:24        Rendered image25    """26    model = YOLO(f'kadirnar/{model_path}-v8.0')27    # set model parameters28    model.overrides['conf'] = conf_threshold  # NMS confidence threshold29    model.overrides['iou'] = iou_threshold  # NMS IoU threshold30    model.overrides['agnostic_nms'] = False  # NMS class-agnostic31    model.overrides['max_det'] = 1000  # maximum number of detections per image32    results = model.predict(image, imgsz=image_size)33    render = render_result(model=model, image=image, result=results[0])34    return render35        36 37inputs = [38    gr.Image(type="filepath", label="Input Image"),39    gr.Dropdown(["yolov8n", "yolov8m", "yolov8l", "yolov8x"], 40                       value="yolov8m", label="Model"),41    gr.Slider(minimum=320, maximum=1280, value=640, step=320, label="Image Size"),42    gr.Slider(minimum=0.0, maximum=1.0, value=0.25, step=0.05, label="Confidence Threshold"),43    gr.Slider(minimum=0.0, maximum=1.0, value=0.45, step=0.05, label="IOU Threshold"),44]45 46outputs = gr.Image(type="filepath", label="Output Image")47title = "State-of-the-Art YOLO Models for Object detection"48 49examples = [['demo_01.jpg', 'yolov8n', 640, 0.25, 0.45], ['demo_02.jpg', 'yolov8l', 640, 0.25, 0.45], ['demo_03.jpg', 'yolov8x', 1280, 0.25, 0.45]]50demo_app = gr.Interface(51    fn=yolov8_inference,52    inputs=inputs,53    outputs=outputs,54    title=title,55    examples=examples,56    cache_examples=True,57)58demo_app.launch(debug=True)