app.py

import gradio as gr
import torch
from PIL import Image
import json

# Images
torch.hub.download_url_to_file(
    'https://iiif.dl.itc.u-tokyo.ac.jp/iiif/genji/TIFF/A00_6587/01/01_0004.tif/full/1024,/0/default.jpg', '『源氏物語』(東京大学総合図書館所蔵).jpg')
torch.hub.download_url_to_file(
    'https://rmda.kulib.kyoto-u.ac.jp/iiif/RB00007030/01/RB00007030_00003_0.ptif/full/1024,/0/default.jpg', '『源氏物語』(京都大学所蔵).jpg')
torch.hub.download_url_to_file(
    'https://kotenseki.nijl.ac.jp/api/iiif/100312034/v4/HRSM/HRSM-00396/HRSM-00396-00012.tif/full/1024,/0/default.jpg', '『平家物語』(国文学研究資料館提供).jpg')

# Model
# model = torch.hub.load('ultralytics/yolov5', 'yolov5s')  # force_reload=True to update
model = torch.hub.load('ultralytics/yolov5', 'custom',
                       path='best.pt', source="local")


def yolo(im, size=1024):
    g = (size / max(im.size))  # gain
    im = im.resize((int(x * g) for x in im.size), resample=Image.Resampling.LANCZOS)  # resize

    results = model(im)  # inference
    results.render()  # updates results.imgs with boxes and labels

    df = results.pandas().xyxy[0].to_json(orient="records")
    res = json.loads(df)

    return [
        Image.fromarray(results.imgs[0]),
        res
    ]


inputs = gr.inputs.Image(type='pil', label="Original Image")
outputs = [
    gr.outputs.Image(type="pil", label="Output Image"),
    gr.outputs.JSON(label="Output JSON")
]

title = "YOLOv5 NDL-DocL Datasets"
description = "YOLOv5 NDL-DocL Datasets Gradio demo for object detection. Upload an image or click an example image to use."
article = "<p style='text-align: center'>YOLOv5 NDL-DocL Datasets is an object detection model trained on the <a href=\"https://github.com/ndl-lab/layout-dataset\">NDL-DocL Datasets</a>.</p>"

examples = [['『源氏物語』(東京大学総合図書館所蔵).jpg'], ['『源氏物語』(京都大学所蔵).jpg'], ['『平家物語』(国文学研究資料館提供).jpg']]
gr.Interface(yolo, inputs, outputs, title=title, description=description, article=article,
             examples=examples, theme="huggingface").launch(enable_queue=True)  # cache_examples=True,