app.py

import spaces
import PIL.Image as Image
import gradio as gr
from ultralytics import ASSETS, YOLO


model = YOLO("best.pt")

@spaces.GPU
def predict_image(img, conf_threshold, iou_threshold):
    results = model.predict(
        source=img,
        device="cuda:0",
        conf=conf_threshold,
        iou=iou_threshold,
        show_labels=True,
        show_conf=True,
        imgsz=416,
        max_det=1
    )

    for r in results:
        im_array = r.plot()
        im = Image.fromarray(im_array[..., ::-1])

    return im


asl = gr.Interface(
    fn=predict_image,
    inputs=[
        gr.Image(type="pil", label="Upload Image"),
        gr.Slider(minimum=0, maximum=1, value=0.25, label="Confidence threshold"),
        gr.Slider(minimum=0, maximum=1, value=0.45, label="IoU threshold")
    ],
    outputs=gr.Image(type="pil", label="Result"),
    title="ASL Detector YOLOv9e",
    description="Upload images for inference. The Ultralytics YOLOv9e model is used by default. Letter Z is not supported.",
    examples=[
        [ASSETS / "a.jpg", 0.25, 0.45],
        [ASSETS / "b.jpg", 0.25, 0.45],
    ]
)

if __name__ == '__main__':
    asl.launch()