|
|
|
from huggingface_hub import InferenceClient |
|
import gradio as gr |
|
|
|
|
|
client = InferenceClient("mistralai/Mixtral-8x7B-Instruct-v0.1") |
|
|
|
|
|
def format_prompt(message, history, system_prompt): |
|
prompt = "<s>" |
|
for user_prompt, bot_response in history: |
|
prompt += f"[INST] {user_prompt} [/INST]" |
|
prompt += f" {bot_response}</s> " |
|
prompt += f"[INST] {system_prompt}, {message} [/INST]" |
|
return prompt |
|
|
|
|
|
def generate( |
|
prompt, history, system_prompt, temperature=0.9, max_new_tokens=4096, top_p=0.95, repetition_penalty=1.0,): |
|
|
|
temperature = float(temperature) |
|
if temperature < 1e-2: |
|
temperature = 1e-2 |
|
top_p = float(top_p) |
|
|
|
|
|
generate_kwargs = dict( |
|
temperature=temperature, |
|
max_new_tokens=max_new_tokens, |
|
top_p=top_p, |
|
repetition_penalty=repetition_penalty, |
|
do_sample=True, |
|
seed=42, |
|
) |
|
|
|
|
|
formatted_prompt = format_prompt(prompt, history, system_prompt) |
|
stream = client.text_generation(formatted_prompt, **generate_kwargs, stream=True, details=True, return_full_text=False) |
|
output = "" |
|
|
|
|
|
for response in stream: |
|
output += response.token.text |
|
yield output |
|
return output |
|
|
|
|
|
additional_inputs = [ |
|
|
|
gr.Textbox( |
|
label="System Prompt", |
|
value="Experto en servicios de abastecimiento, depuracion, reutilizacion y calidad del agua, para la empresa Canal de Isabel II", |
|
max_lines=1, |
|
interactive=True, |
|
), |
|
|
|
gr.Slider( |
|
label="Temperature", |
|
value=0.9, |
|
minimum=0.0, |
|
maximum=1.0, |
|
step=0.05, |
|
interactive=True, |
|
info="Valores más altos producen resultados más diversos", |
|
), |
|
|
|
|
|
gr.Slider( |
|
label="Max new tokens", |
|
value=4096, |
|
minimum=0, |
|
maximum=4096, |
|
step=64, |
|
interactive=True, |
|
info="El máximo número de nuevos tokens", |
|
), |
|
|
|
gr.Slider( |
|
label="Top-p (nucleus sampling)", |
|
value=0.90, |
|
minimum=0.0, |
|
maximum=1, |
|
step=0.05, |
|
interactive=True, |
|
info="Valores más altos muestrean más tokens de baja probabilidad", |
|
), |
|
|
|
gr.Slider( |
|
label="Repetition penalty", |
|
value=1.2, |
|
minimum=1.0, |
|
maximum=2.0, |
|
step=0.05, |
|
interactive=True, |
|
info="Penaliza los tokens repetidos", |
|
) |
|
] |
|
|
|
|
|
examples = [ |
|
["que parametros definen la calidad del agua", "Experto en servicios de abastecimiento, depuracion, reutilizacion y calidad del agua, para la empresa Canal de Isabel II", 0.7, 150, 0.80, 1.1], |
|
["Describeme el proceso basico de depuracion del agua", "Experto en servicios de abastecimiento, depuracion y reutilizacion del agua", 0.8, 250, 0.85, 1.2], |
|
["¿Que es una acometida?", "Experto en servicios de abastecimiento, depuracion, reutilizacion y calidad del agua, para la empresa Canal de Isabel II", 0.7, 180, 0.75, 1.2], |
|
["¿Que es una toma?", "Experto en servicios de abastecimiento, depuracion, reutilizacion y calidad del agua, para la empresa Canal de Isabel II", 0.8, 180, 0.80, 1.1], |
|
] |
|
|
|
|
|
gr.ChatInterface( |
|
fn=generate, |
|
chatbot=gr.Chatbot(show_label=False, show_share_button=False, show_copy_button=True, likeable=True, layout="panel"), |
|
additional_inputs=additional_inputs, |
|
title="Mixtral 8B Fines didácticos", |
|
description='Autor: <a href=\"https://huggingface.co/Antonio49\">Antonio Fernández</a> de <a href=\"https://saturdays.ai/\">SaturdaysAI</a>. Formación: <a href=\"https://cursos.saturdays.ai/courses/\">Cursos Online AI</a> Aplicación desarrollada con fines docentes', |
|
theme="soft", |
|
examples=examples, |
|
cache_examples=True, |
|
retry_btn="Repetir", |
|
undo_btn="Deshacer", |
|
clear_btn="Borrar", |
|
submit_btn="Enviar", |
|
concurrency_limit=20, |
|
).launch(show_api=False) |
|
|