from transformers import pipeline
import gradio as gr
import cv2
from PIL import Image


image_to_text = pipeline("image-to-text", model="nlpconnect/vit-gpt2-image-captioning")


def ocr(image):
    inp = Image.fromarray(image.astype('uint8'), 'RGB')
    text = image_to_text(inp)

    total_caption = "" 
    for caption in text:
        total_caption = total_caption + caption.get('generated_text')
        total_caption = total_caption + '\r\n'

    return total_caption 

demo = gr.Interface(fn=ocr,
                    inputs='image',
                    outputs='text',
                    title = "image2text",
                    examples = ['soccer.jpg'])


if __name__ == "__main__":
    demo.queue(concurrency_count=3).launch(server_name = "0.0.0.0", server_port = 7016)