Update app.py

app.py

@@ -1,18 +1,11 @@
 import streamlit as st
 from transformers import AutoTokenizer, AutoModelForCausalLM
 import torch
-import subprocess
-from dotenv import load_dotenv
-import os
+
 # Function to load model and tokenizer only once
 @st.cache_resource  # This decorator caches the loading process
 def load_resources():
-    load_dotenv()
     # Load the pre-trained Llama3 model (or your fine-tuned model)
-    huggingface_token = os.getenv("HUGGINGFACE_TOKEN")
-
-    # Run the huggingface-cli login command from the Python script using subprocess
-    subprocess.run(["huggingface-cli", "login", "--token", huggingface_token])
     tokenizer = AutoTokenizer.from_pretrained("meta-llama/Llama-3.2-1B")
     model = AutoModelForCausalLM.from_pretrained("meta-llama/Llama-3.2-1B")
     print("Model and tokenizer loaded.")
@@ -34,6 +27,10 @@ user_query = st.text_input("Your Query:", "")
 
 if user_query:
     with st.spinner("Generating response..."):
+        # Retrieve the tokenizer from session state
+        tokenizer = st.session_state.tokenizer
+        model = st.session_state.model
+        
         # Tokenize the input query
         inputs = tokenizer(user_query, return_tensors="pt")