Update app.py

app.py

@@ -3,7 +3,7 @@ import torch
 from sentence_transformers import SentenceTransformer, util
 import gradio as gr
 import json
-from transformers import AutoTokenizer, AutoModelForCausalLM
+from transformers import AutoTokenizer, AutoModelForQuestionAnswering
 import spaces
 
 # Ensure you have GPU support
@@ -20,8 +20,8 @@ embeddings = torch.tensor(df['embedding'].tolist(), device=device)
 model = SentenceTransformer('all-MiniLM-L6-v2', device=device)
 
 # Load the LLaMA model for response generation
-llama_tokenizer = AutoTokenizer.from_pretrained("openai-community/gpt2")
-llama_model = AutoModelForCausalLM.from_pretrained("openai-community/gpt2").to(device)
+llama_tokenizer = AutoTokenizer.from_pretrained("distilbert/distilbert-base-uncased-distilled-squad")
+llama_model = AutoModelForQuestionAnswering.from_pretrained("distilbert/distilbert-base-uncased-distilled-squad").to(device)
 
 # Define the function to find the most relevant document
 @spaces.GPU(duration=120)