Fix tokenizer and import issues for T5 Indonesian model

app.py

@@ -1,13 +1,12 @@
 import streamlit as st
 from newspaper import Article
-from transformers import pipeline, AutoTokenizer, AutoModelForSeq2SeqLM
+from transformers import pipeline, T5Tokenizer, AutoModelForSeq2SeqLM
 
 # Load model from Hugging Face
 @st.cache_resource
 def load_summarizer():
     model_name = "cahya/t5-base-indonesian-summarization-cased"
-    
-    tokenizer = AutoTokenizer.from_pretrained(model_name, use_fast=False)
+    tokenizer = T5Tokenizer.from_pretrained(model_name)
     model = AutoModelForSeq2SeqLM.from_pretrained(model_name)
 
     return pipeline("summarization", model=model, tokenizer=tokenizer)

requirements.txt

@@ -1,5 +1,7 @@
 streamlit
 newspaper3k
-transformers
+transformers==4.36.2
 torch
+tiktoken
+sentencepiece
 lxml[html_clean]