| import gradio as gr | |
| import subprocess | |
| # Define the inference function using llama.cpp | |
| def predict(text): | |
| # Call llama.cpp with the input text | |
| result = subprocess.run( | |
| ["./llama.cpp/main", "-m", "GSridhar1982/QA_Llama31_Quantized_GGUF", "-p", text], | |
| capture_output=True, | |
| text=True | |
| ) | |
| return result.stdout | |
| # Create a Gradio interface | |
| iface = gr.Interface( | |
| fn=predict, | |
| inputs=gr.Textbox(lines=2, placeholder="Enter question here..."), | |
| outputs="Answer", | |
| title="LLaMA Model Inference", | |
| description="Enter text to generate using the LLaMA model." | |
| ) | |
| # Launch the interface | |
| iface.launch() | |