update app.py

app.py

@@ -1,4 +1,3 @@
-# Import necessary libraries
 import streamlit as st
 import numpy as np
 from transformers import AutoModelForSeq2SeqLM, AutoTokenizer
@@ -6,30 +5,26 @@ from normalizer import normalize
 
 # Set the page configuration
 st.set_page_config(
-    page_title="Bengalai to English Translator App",  # Title of the app displayed in the browser tab
+    page_title="Bengali to English Translator App",  # Title of the app displayed in the browser tab
     page_icon=":shield:",  # Path to a favicon or emoji to be displayed in the browser tab
     initial_sidebar_state="auto"  # Initial state of the sidebar ("auto", "expanded", or "collapsed")
 )
 
-
 # Load custom CSS styling
 with open("assets/style.css") as f:
     st.markdown("<style>{}</style>".format(f.read()), unsafe_allow_html=True)
 
 # Function to load the pre-trained model
-# @st.cache_data(experimental_allow_widgets=False)
 def get_model():
-    tokenizer = AutoTokenizer.from_pretrained("kazalbrur/BanglaEnglishTokenizerBanglaT5", use_fast=True) # Set legacy=False
-    model = AutoModelForSeq2SeqLM.from_pretrained("kazalbrur/BanglaEnglishTranslationBanglaT5") # Set legacy=False
+    tokenizer = AutoTokenizer.from_pretrained("kazalbrur/BanglaEnglishTokenizerBanglaT5", use_fast=True)
+    model = AutoModelForSeq2SeqLM.from_pretrained("kazalbrur/BanglaEnglishTranslationBanglaT5")
     return tokenizer, model
 
-
 # Load the tokenizer and model
 tokenizer, model = get_model()
 
-
 # Add a header to the Streamlit app
-st.header("Benglai to English Translator")
+st.header("Bengali to English Translator")
 
 # Add placeholder text with custom CSS styling
 st.markdown("<span style='color:black'>Enter your Bengali text here</span>", unsafe_allow_html=True)
@@ -43,9 +38,6 @@ submit_button = st.button("Translate")
 # Perform prediction when user input is provided and the submit button is clicked
 if user_input and submit_button:
     input_ids = tokenizer(normalize(user_input), padding=True, truncation=True, max_length=128, return_tensors="pt").input_ids
-    generated_tokens = model.generate(input_ids, max_new_tokens=128)  # Set max_new_tokens to control generation length
+    generated_tokens = model.generate(input_ids, max_new_tokens=128)
     decoded_tokens = tokenizer.batch_decode(generated_tokens, skip_special_tokens=True)[0]
-    st.write(f"<span style='color:black'>Bangla Translation: {decoded_tokens}</span>", unsafe_allow_html=True)
-
-
-
+    st.write(f"<span style='color:black'>English Translation: {decoded_tokens}</span>", unsafe_allow_html=True)