app.py

import streamlit as st

st.title("Grammar Corrector")
st.write("Paste or type text, submit and the machine will attempt to correct your text's grammar.")

default_text = " Lastly, another reason we will need clear procedures so that customers do not exploit our refund policy."
sent = st.text_area("Text", default_text, height=40)
num_correct_options = st.number_input('Number of Correction Options', min_value=1, max_value=3, value=1, step=1)

from transformers import T5ForConditionalGeneration, T5Tokenizer
import torch
torch_device = 'cuda' if torch.cuda.is_available() else 'cpu'
tokenizer = T5Tokenizer.from_pretrained('deep-learning-analytics/GrammarCorrector')
model = T5ForConditionalGeneration.from_pretrained('deep-learning-analytics/GrammarCorrector').to(torch_device)

def correct_grammar(input_text, num_return_sequences=num_correct_options):
  batch = tokenizer([input_text], truncation=True, padding = 'max_length', max_length = 64, return_tensors = 'pt').to(torch_device)
  results = model.generate(**batch, max_length = 64, num_beams = 4, num_return_sequences = num_correct_options, temperature = 1.5)
  
  return results
  
results = correct_grammar(sent, num_correct_options)

generated_options = []
for generated_option_idx, generated_option in enumerate(results):
    
    text = tokenizer.decode(generated_option, clean_up_tokenization_spaces = True, skip_special_tokens = True)
    generated_options.append(text)
    
st.write(generated_options)