app.py

import gradio as gr
import tempfile
from TTS.utils.synthesizer import Synthesizer
from huggingface_hub import hf_hub_download

REPO_ID = "mbarnig/lb-de-fr-en-pt-coqui-vits-tts"

my_title = "🇩🇪 🇫🇷 🇬🇧 🇵🇹 Mir schwätzen och Lëtzebuergesch ! 🇱🇺"
my_description = "First multilingual-multispeaker Text-to-Speech (TTS) synthesizer speaking Luxembourgish. This model is based on [YourTTS](https://github.com/Edresson/YourTTS), thanks to 🐸 [Coqui.ai](https://coqui.ai/)." 
lb_text = "An der Zäit hunn sech den Nordwand an d'Sonn gestridden, wie vun hinnen zwee wuel méi staark wier, wéi e Wanderer, deen an ee waarme Mantel agepak war, iwwert de Wee koum."
de_text = "Einst stritten sich Nordwind und Sonne, wer von ihnen beiden wohl der Stärkere wäre, als ein Wanderer, der in einen warmen Mantel gehüllt war, des Weges daherkam."
fr_text = "La bise et le soleil se disputaient, chacun assurant qu'il était le plus fort, quand ils ont vu un voyageur qui s'avançait, enveloppé dans son manteau."
en_text = "The North Wind and the Sun were disputing which was the stronger, when a traveler came along wrapped in a warm cloak."
pt_text = "O vento norte e o Sol discutiam quem era o mais forte, quando surgiu um viajante envolvido numa capa."

TTS_VOICES = [
  "Bernard",
  "Bunny",
  "Ed",
  "Guy",
  "Judith",
  "Kerstin",
  "Linda",
  "Thorsten"
]  

TTS_LANGUAGES = [
  "Deutsch",
  "English",
  "Français",
  "Lëtzebuergesch",  
  "Português"
]

my_examples = [
  [lb_text, "Judith", "Lëtzebuergesch"],
  [de_text, "Thorsten", "Deutsch"],
  [fr_text, "Bernard", "Français"],
  [en_text, "Linda", "English"],
  [pt_text, "Ed", "Português"]
]

my_article = "<h3>User guide</h3><p>1. Press the Submit button to generate a speech file with the default values. 2. Change the default values by clicking an example row. 3. Select a language and a voice and enter your own text. Have fun!</p><p>Go to <a href='https://www.web3.lu/category/audio-technologies/'>Internet with a Brain</a> to read some technical infos.</p>"

my_inputs = [
  gr.Textbox(lines=5, label="Input Text"),
  gr.Radio(label="Speaker", choices = TTS_VOICES, value = "Judith"),
  gr.Radio(label="Language", choices = TTS_LANGUAGES, value = "Lëtzebuergesch"),
]

my_outputs = gr.Audio(type="filepath", label="Output Audio")

def tts(text: str, speaker_idx: str, language_idx: str):
    best_model_path = hf_hub_download(repo_id=REPO_ID, filename="best_model.pth") 
    config_path = hf_hub_download(repo_id=REPO_ID, filename="config.json")
    speakers_path = hf_hub_download(repo_id=REPO_ID, filename="speakers.pth")
    languages_path = hf_hub_download(repo_id=REPO_ID, filename="language_ids.json")
    speaker_encoder_model_path = hf_hub_download(repo_id=REPO_ID, filename="model_se.pth")
    speaker_encoder_config_path = hf_hub_download(repo_id=REPO_ID, filename="config_se.json")
    
    # init synthesizer  
    synthesizer = Synthesizer(
        best_model_path,
        config_path,
        speakers_path,
        languages_path,
        None,
        None,
        speaker_encoder_model_path,
        speaker_encoder_config_path,
        False
    )

    # create audio file
    wavs = synthesizer.tts(text, speaker_idx, language_idx)
    with tempfile.NamedTemporaryFile(suffix = ".wav", delete = False) as fp:
        synthesizer.save_wav(wavs, fp)                      
    return fp.name 
 
iface = gr.Interface(
    fn=tts, 
    inputs=my_inputs, 
    outputs=my_outputs, 
    title=my_title, 
    description = my_description, 
    article = my_article,
    examples = my_examples, 
    allow_flagging=False
)
iface.launch()