app.py

import streamlit as st

from transformers import TrOCRProcessor, VisionEncoderDecoderModel
from PIL import Image

processor = TrOCRProcessor.from_pretrained("microsoft/trocr-base-handwritten")
model = VisionEncoderDecoderModel.from_pretrained("microsoft/trocr-base-handwritten")

def process_image(image):
  # prepare image
  pixel_values = processor(image, return_tensors="pt").pixel_values
  # generate (no beam search)
  generated_ids = model.generate(pixel_values)
  # decode
  generated_text = processor.batch_decode(generated_ids, skip_special_tokens=True)[0]
  return generated_text

########################## Streamlit Code ##########################
st.title('Streamlit Replication of nielsr/TrOCR-handwritten')
uploaded_file = st.file_uploader("Choose an image...")
if uploaded_file:
  # .convert('RGB') to mode=RGB
  input_image = Image.open(uploaded_file).convert('RGB')
  st.image(uploaded_file, caption='Input Image', use_column_width=True)
  generated_text = process_image(input_image)
  st.write(generated_text)