app.py

import gradio as gr
from transformers import AutoModelForCausalLM, AutoTokenizer, TextIteratorStreamer
from threading import Thread

model_id = "AuriLab/gpt-bi"
tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(model_id)

def generar_texto(prompt):
    streamer = TextIteratorStreamer(tokenizer, skip_prompt=True)
    
    inputs = tokenizer(prompt, return_tensors="pt")
    
    thread = Thread(target=model.generate, kwargs={
        "input_ids": inputs.input_ids,
        "max_new_tokens": 50,
        "streamer": streamer,
        "do_sample": True
    })
    thread.start()
    
    output = prompt
    yield output
    
    for new_token in streamer:
        output += new_token
        yield output

textbox = gr.Textbox(
    lines=5,
    placeholder="Escribe tu prompt aquí...",
    label="Generación en tiempo real"
)


interfaz = gr.Interface(
    fn=generar_texto,
    inputs=textbox,
    outputs=textbox,
    title="Generador de Texto con Streamer",
    description="Escribe y los tokens aparecerán en tiempo real en la misma caja"
)

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