Update app.py

app.py

@@ -1,10 +1,10 @@
 import streamlit as st
 import fitz  # PyMuPDF
 # Load model directly
-from transformers import AutoTokenizer, AutoModelForQuestionAnswering
+from transformers import AutoTokenizer, AutoModelForCausalLM
 
-tokenizer = AutoTokenizer.from_pretrained("google-bert/bert-large-cased-whole-word-masking-finetuned-squad")
-model = AutoModelForQuestionAnswering.from_pretrained("google-bert/bert-large-cased-whole-word-masking-finetuned-squad")
+tokenizer = AutoTokenizer.from_pretrained("google/bert_uncased_L-12_H-768_A-12")
+model = AutoModelForCausalLM.from_pretrained("google/bert_uncased_L-12_H-768_A-12")
 
 # Function to extract text from PDF
 def extract_text_from_pdf(pdf_file):
@@ -20,7 +20,6 @@ def generate_mcqs(text, num_questions=5):
     if not text.strip():
         return ["No text extracted from the PDF. Unable to generate MCQs."]
 
-
     # Tokenize the input text
     inputs = tokenizer(text, return_tensors="pt", max_length=512, truncation=True)
 
@@ -30,7 +29,6 @@ def generate_mcqs(text, num_questions=5):
         output = model.generate(
             **inputs,
             max_length=512,  # Adjusted to match the tokenized input length
-            max_new_tokens=100,  # Limit the number of new tokens added during generation
             temperature=0.7,
             num_return_sequences=1
         )