app.py

import gradio as gr  
import subprocess  

from llama_cpp import Llama

# Initialize the model
model = Llama(model_path="QA_Llama31_Quantized_GGUF")

def generate_response(prompt):
    response = model.create_chat_completion(messages=[{"role": "user", "content": prompt}])
    return response['choices'][0]['message']['content']

# Define the inference function using llama.cpp  
def predict(text):  
    # Call llama.cpp with the input text  
    result = subprocess.run(  
        ["./llama.cpp/main", "-m", "QA_Llama31_Quantized_GGUF", "-p", text],  
        capture_output=True,  
        text=True  
    )  
    return result.stdout  

# Create a Gradio interface  
iface = gr.Interface(  
    fn=generate_response,  
    inputs=gr.Textbox(lines=2, placeholder="Enter question here..."),  
    outputs="Answer",  
)  

# Launch the interface  
iface.launch()