app.py

import gradio as gr
from huggingface_hub import InferenceClient
import spaces

# Importar spaces para ZeroGPU
client = InferenceClient("xezpeleta/latxa-7b-v1-gguf")

@spaces.GPU  # Añadir el decorador para ZeroGPU
def respond(
    message,
    history: list[tuple[str, str]],
    system_message,
    max_tokens,
    temperature,
    top_p,
):
    messages = [{"role": "system", "content": system_message}]

    for val in history:
        if val[0]:
            messages.append({"role": "user", "content": val[0]})
        if val[1]:
            messages.append({"role": "assistant", "content": val[1]})

    messages.append({"role": "user", "content": message})

    response = ""

    for message in client.chat_completion(
        messages,
        max_tokens=max_tokens,
        stream=True,
        temperature=temperature,
        top_p=top_p,
    ):
        token = message.choices[0].delta.content
        response += token
        yield response

# Configurar la interfaz
demo = gr.ChatInterface(
    respond,
    additional_inputs=[
        gr.Textbox(
            value="Euskarazko laguntzaile adimentsua zara.", 
            label="Sistema mezua"
        ),
        gr.Slider(
            minimum=1, 
            maximum=2048, 
            value=512, 
            step=1, 
            label="Token maximoak"
        ),
        gr.Slider(
            minimum=0.1, 
            maximum=1.0, 
            value=0.7, 
            step=0.1, 
            label="Temperatura"
        ),
        gr.Slider(
            minimum=0.1,
            maximum=1.0,
            value=0.95,
            step=0.05,
            label="Top-p"
        ),
    ],
    title="Latxa LLM Chatbot"
)

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