app.py

# Inference

import gradio as gr
from huggingface_hub import InferenceClient

model = "meta-llama/Llama-3.2-3B-Instruct"
client = InferenceClient(model)

def fn(
    prompt,
    history,
    system_prompt,
    max_tokens,
    temperature,
    top_p,
):
    messages = [{"role": "system", "content": system_prompt}]
    messages.append({"role": "user", "content": prompt})

    history = []
    history.append(messages)
    
    #messages = [{"role": "user", "content": prompt}]
    history.append(messages[0])

    stream = client.chat.completions.create(
        model = model,
        messages = history,
        max_tokens = max_tokens,
        temperature = temperature,
        top_p = top_p,
        stream = True
    )

    chunks = []
    for chunk in stream:
        chunks.append(chunk.choices[0].delta.content or "")
        yield "".join(chunks)

app = gr.ChatInterface(
    fn = fn,
    type = "messages",
    additional_inputs = [
        gr.Textbox(value="You are a helpful assistant.", label="System Prompt"),
        gr.Slider(minimum=1, maximum=2048, value=512, step=1, label="Max Tokens"),
        gr.Slider(minimum=0.1, maximum=4.0, value=0.7, step=0.1, label="Temperature"),
        gr.Slider(minimum=0.1, maximum=1.0, value=0.95, step=0.05, label="Top-P"),
    ],
    title = "Meta Llama",
    description = model,
)

if __name__ == "__main__":
    app.launch()