Update app.py

app.py

@@ -1,28 +1,27 @@
 import streamlit as st
-from transformers import pipeline, T5Tokenizer, T5ForConditionalGeneration
-
-from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
+from transformers import pipeline, AutoTokenizer, AutoModelForSeq2SeqLM
 
 model_repo_path = 'nxmwxm/correct_answer'
 
-
+# Load the tokenizer and model
+tokenizer = AutoTokenizer.from_pretrained(model_repo_path)
+model = AutoModelForSeq2SeqLM.from_pretrained(model_repo_path)
 
 # Load the pipeline with your model and tokenizer
 qa_pipeline = pipeline(
     'question-answering',
-    model=model_repo_path,
-    tokenizer=model_repo_path
+    model=model,
+    tokenizer=tokenizer
 )
 
-def generate_answer(question, distractors):
-    input_text = f"Question: {question} Distractors: {' '.join(distractors)}"
-    result = qa_pipeline(question=input_text, context=input_text)
-    return result['answer']
+def generate_answer(question, context):
+    result = qa_pipeline(question=question, context=context)
+    return result.get('answer', 'No answer found')
 
 st.title('Question Answering Model')
 question = st.text_input('Enter your question:')
-distractors = st.text_input('Enter distractors (comma separated):').split(',')
+context = st.text_area('Enter context for the question:')
 
 if st.button('Generate Answer'):
-    answer = generate_answer(question, distractors)
+    answer = generate_answer(question, context)
     st.write('Answer:', answer)