app.py

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

# Load PaddleOCR-VL model
model_name = "PaddlePaddle/PaddleOCR-VL"
tokenizer = AutoTokenizer.from_pretrained(model_name, trust_remote_code=True)
model = AutoModel.from_pretrained(model_name, trust_remote_code=True)

if torch.cuda.is_available():
    model = model.cuda()

@spaces.GPU
def ocr_inference(image):
    """
    Perform OCR on the input image using PaddleOCR-VL
    """
    if image is None:
        return "Please upload an image."
    
    try:
        # Convert to PIL Image if needed
        if not isinstance(image, Image.Image):
            image = Image.fromarray(image)
        
        # Run OCR inference
        result = model.chat(tokenizer, image, "Extract all text from this image.")
        return result
    except Exception as e:
        return f"Error during OCR: {str(e)}"

# Create Gradio interface
demo = gr.Interface(
    fn=ocr_inference,
    inputs=gr.Image(type="pil", label="Upload Image for OCR"),
    outputs=gr.Textbox(label="Extracted Text"),
    title="PaddleOCR-VL OCR Demo",
    description="Upload an image to extract text using PaddlePaddle/PaddleOCR-VL model"
)

demo.launch()