app.py

import gradio as gr
import numpy as np
import torch
from huggingsound import SpeechRecognitionModel
from diffusers import StableDiffusionPipeline, DPMSolverMultistepScheduler
from transformers import pipeline

# Función para convertir la tasa de muestreo del audio de entrada
def modelo1(audio):
    print(audio)
    whisper = pipeline('automatic-speech-recognition', model='openai/whisper-medium', device=-1)  # Cambia 'device' a -1 para usar la CPU
    print(np.array(audio[1]))
    text = whisper(np.array(audio[1]))
    return text["text"]
def modelo2(text):
    model_id = "stabilityai/stable-diffusion-2-1"

    # Use the DPMSolverMultistepScheduler (DPM-Solver++) scheduler here instead
    pipe = StableDiffusionPipeline.from_pretrained(model_id, torch_dtype=torch.float16)
    pipe.scheduler = DPMSolverMultistepScheduler.from_config(pipe.scheduler.config)
    pipe = pipe.to("cpu")
    print(text)
 # Conversión a float32
    image = pipe(text).images[0]

    return image

def execution(audio):
    modelo1res = modelo1(audio)
    modelo2res = modelo2(modelo1res)
    return modelo2res

if __name__ == "__main__":
    demo = gr.Interface(fn=execution, inputs="audio", outputs="image")
    demo.launch()