app.py

import streamlit as st
from transformers import AutoModelForCausalLM, AutoTokenizer
import torch

# Load the model and tokenizer
@st.cache_resource
def load_model():
    model_name = "prithivMLmods/QwQ-LCoT-14B-Conversational"
    tokenizer = AutoTokenizer.from_pretrained(model_name)
    model = AutoModelForCausalLM.from_pretrained(
        model_name,
        device_map="auto",  # Automatically assign to CPU/GPU
        torch_dtype=torch.float16,  # Mixed precision for large models
    )
    return tokenizer, model

# Load resources
tokenizer, model = load_model()

# Streamlit app UI
st.title("QwQ-LCoT Chatbot")
st.write("A conversational AI powered by QwQ-LCoT-14B. Ask me anything!")

# User input
user_input = st.text_input("You: ", "")

if st.button("Send"):
    if user_input.strip():
        with st.spinner("Generating response..."):
            # Tokenize input
            inputs = tokenizer(user_input, return_tensors="pt")
            # Generate response
            outputs = model.generate(**inputs, max_new_tokens=150, temperature=0.7)
            # Decode response
            response = tokenizer.decode(outputs[0], skip_special_tokens=True)
            
        # Display response
        st.text_area("Bot:", value=response, height=150)