app.py

import gradio as gr
import torch
from PIL import Image

from donut import DonutModel

def demo_process(input_img):
    global pretrained_model, task_prompt, task_name
    # input_img = Image.fromarray(input_img)
    output = pretrained_model.inference(image=input_img, prompt=task_prompt)["predictions"][0]
    return output
task_name = "preparedFinetuneData"
# task_name = "cord-v2"
task_prompt = f"<s_{task_name}>"

image = Image.open("sample_receipt1.png")
image.save("sample_receipt1.png")
image = Image.open("sample_receipt2.png")
image.save("sample_receipt2.png")

PATH = 'epochs30_base_on_donut_base/'
# pretrained_model = DonutModel.from_pretrained(PATH, local_files_only=True)
# pretrained_model = DonutModel.from_pretrained("doshan1250/p9OcrAiV1", revision="main")
pretrained_model = DonutModel.from_pretrained("doshan1250/p9OcrAiV1")
# pretrained_model = DonutModel.from_pretrained("naver-clova-ix/donut-base-finetuned-cord-v2")
pretrained_model.eval()

demo = gr.Interface(
    fn=demo_process,
    inputs= gr.Image(type="pil"),
    outputs="json",
    title=f"Donut 🍩 demonstration for `{task_name}` task",
    description="""Goodarc p9 使用 100 個英文收據訓練. <br>""",
    examples=[["sample_receipt1.png"], ["sample_receipt2.png"]],
    cache_examples=False,
)

demo.launch()