app.py

import streamlit as st
from huggingface_hub import InferenceClient

# Access your Hugging Face token (stored as a secret in the environment)
hf_token = st.secrets["Hf_Token"]

# Initialize the Hugging Face Inference Client with the token
client = InferenceClient(token=hf_token)

# Streamlit App UI
st.title("LLM Model Inference with Streamlit")

# Input text box for user prompt
user_input = st.text_input("Enter a text prompt:", "")

# Dropdown for selecting between LLaMA 3 models
model_id = st.selectbox(
    "Select a LLaMA model",
    ["meta-llama/Meta-Llama-3-2B", "meta-llama/Meta-Llama-3-3B"]  # Replace with the correct model names
)

# Button to trigger the inference
if st.button("Generate Text"):
    if user_input:
        with st.spinner(f"Generating text using {model_id}..."):
            # Perform inference using the selected model
            response = client.text_generation(model=model_id, inputs=user_input)
            st.success("Text generated!")
            st.write(response)
    else:
        st.warning("Please enter a prompt to generate text.")