xtts-spanish / app.py
mrm8488's picture
Update app.py
1546648
raw history blame
No virus
2.23 kB
import gradio as gr
from TTS.api import TTS
model_id = "tts_models/multilingual/multi-dataset/xtts_v1"
device = "cuda"
tts = TTS(model_id)
tts.to(device)
def predict(prompt, audio_file_pth):
tts.tts_to_file(
text=prompt,
file_path="output.wav",
speaker_wav=audio_file_pth,
language=language,
)
return (
gr.make_waveform(
audio="output.wav",
),
"output.wav",
)
title = "Coquib🐸 XTTS - Spanish Demo"
description = """
<a href="https://huggingface.co/coqui/XTTS-v1">XTTS</a> is a Voice generation model that lets you clone voices into different languages by using just a quick 3-second audio clip.
<br/>
Built on Tortoise, XTTS has important model changes that make cross-language voice cloning and multi-lingual speech generation super easy.
<br/>
This is the same model that powers Coqui Studio, and Coqui API, however we apply a few tricks to make it faster and support streaming inference.
<br/>
<br/>
<p>For faster inference without waiting in the queue, you should duplicate this space and upgrade to GPU via the settings.
<br/>
<a href="https://huggingface.co/spaces/coqui/xtts?duplicate=true">
<img style="margin-top: 0em; margin-bottom: 0em" src="https://bit.ly/3gLdBN6" alt="Duplicate Space"></a>
</p>
"""
article = """
<div style='margin:20px auto;'>
<p>By using this demo you agree to the terms of the Coqui Public Model License at https://coqui.ai/cpml</p>
</div>
"""
gr.Interface(
fn=predict,
inputs=[
gr.Textbox(
label="Texto",
info="Una o dos frases es suficiente-",
value="Clibrain es una empresa que desarrolla soluciones basadas en inteligencia artificial en español.",
),
gr.Audio(
label="Audio de referencia",
info="Haz clic en el botón ✎ para subir tu propio audio o del hablante objetivo",
type="filepath",
value="examples/female.wav",
),
],
outputs=[
gr.Video(label="Waveform Visual"),
gr.Audio(label="Synthesised Audio"),
],
title=title,
description=description,
article=article,
examples=examples,
).queue().launch(debug=True)