Spaces:
Sleeping
Sleeping
import os | |
import gradio as gr | |
import torch | |
from transformers import AutoModelForCausalLM, AutoTokenizer | |
from huggingface_hub import login | |
# Obtener el token directamente del secreto de Hugging Face | |
huggingface_token = os.environ.get('reparfinal') | |
if huggingface_token is None: | |
raise ValueError("El token de Hugging Face no está configurado en los secretos del Space") | |
# Iniciar sesión | |
login(huggingface_token) | |
# Configurar modelo | |
model_id = "meta-llama/Llama-2-7b-chat-hf" # Cambiado al modelo de 7B | |
tokenizer = AutoTokenizer.from_pretrained(model_id) | |
model = AutoModelForCausalLM.from_pretrained( | |
model_id, | |
torch_dtype=torch.bfloat16, | |
device_map="auto" | |
) | |
def respond_to_query(user_input): | |
# Formato específico para Llama 2 Chat | |
prompt = f"[INST] {user_input} [/INST]" | |
inputs = tokenizer.encode(prompt, return_tensors="pt").to(model.device) | |
outputs = model.generate( | |
inputs, | |
max_new_tokens=256, | |
do_sample=True, | |
top_p=0.95, | |
top_k=50, | |
temperature=0.7, | |
repetition_penalty=1.1 | |
) | |
return tokenizer.decode(outputs[0], skip_special_tokens=True) | |
# Interfaz Gradio | |
interface = gr.Interface( | |
fn=respond_to_query, | |
inputs=gr.Textbox(label="Tu pregunta"), | |
outputs=gr.Textbox(label="Respuesta"), | |
title="Chatbot con Llama-2-7b", | |
description="Haz una pregunta y el modelo te responderá" | |
) | |
if __name__ == "__main__": | |
interface.launch() | |