app.py

import gradio as gr
import torch

#############

def yolov7_inference(
    image: gr.Image = None,
    conf_threshold: gr.Slider = 0.20,
):

    device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")
    path = 'y7-prdef.pt'
    model = torch.hub.load("WongKinYiu/yolov7","custom",f"{path}")
    model.conf = conf_threshold
    results = model([image], size=640)
    return results.render()[0]
        
inputs = [
    gr.Image(type="filepath", label="Input"),
    gr.Slider(minimum=0.0, maximum=1.0, value=0.2, step=0.05, label="Confidence Threshold", interactive=True),
]

outputs = [
    gr.Image(type="filepath"),
    
]

css = ".output_image {height: 40rem !important; width: 100% !important;}"

demo = gr.Interface(
    fn=yolov7_inference,
    inputs=inputs,
    outputs=outputs,
    title="The detection of jar lid defects using Yolov7",
    description = """
This application is detecting damaged jar lids. Type of damages include deformations, holes or scratches. The object detection notebook can be found at <a href="https://www.kaggle.com/rrighart">Kaggle</a>

Contact: Ruthger Righart

Email: [email protected]

Web: <a href="https://www.rrighart.com">www.rrighart.com</a> 
""",
    article = "<p style='text-align: center'><a href='https://www.rrighart.com' target='_blank'>Webpage</a></p> <p style='text-align: center'><a href='https://www.kaggle.com/code/rrighart/detection-of-product-defects-using-yolov7' target='_blank'>Kaggle</a></p>",
    examples = [['example1.JPG'], ['example2.JPG'], ['example3.JPG']],
    #examples = [['example1.JPG', 0.50], ['example2.JPG', 0.50], ['example3.JPG', 0.50]],
    css=css,
    cache_examples=True,
)


demo.queue().launch(server_name="0.0.0.0", server_port=7860, debug=False, inline=True)