app.py.bak

import gradio as gr
from transformers import AutoModel, AutoTokenizer
from PIL import Image
import torch

# Load model and tokenizer
tokenizer = AutoTokenizer.from_pretrained('ucaslcl/GOT-OCR2_0', trust_remote_code=True)
model = AutoModel.from_pretrained(
    'ucaslcl/GOT-OCR2_0',
    trust_remote_code=True,
    low_cpu_mem_usage=True,
    device_map='cuda' if torch.cuda.is_available() else 'cpu',
    use_safetensors=True,
    pad_token_id=tokenizer.eos_token_id
)
model = model.eval()
if torch.cuda.is_available():
    model = model.cuda()

# OCR function
def ocr_from_image(image, ocr_type):
    image_path = "temp.jpg"
    image.save(image_path)
    res = model.chat(tokenizer, image_path, ocr_type=ocr_type)
    return res

# Gradio interface
ocr_types = ["ocr", "format"]

iface = gr.Interface(
    fn=ocr_from_image,
    inputs=[
        gr.Image(type="pil", label="Upload Image"),
        gr.Radio(ocr_types, label="OCR Type", value="ocr")
    ],
    outputs="text",
    title="GOT-OCR2.0: OCR with Transformers",
    description="Upload an image and select OCR type (plain text or formatted)."
)

if __name__ == "__main__":
    iface.launch()