app.py

import os

os.system("git clone --branch v2.3 https://github.com/DigitalPhonetics/IMS-Toucan.git toucan_codebase")
os.system("mv toucan_codebase/* .")

from run_model_downloader import download_models
from run_gradio_demo import TTSWebUI

download_models()
TTSWebUI(gpu_id="cpu", title="Design a Speaker", article="We use a GAN to generate a speaker embedding for the TTS to use. The GAN has a control mechanism integrated, so you can directly influence the generated embeddings in multiple intuitive ways! When combining this with a bunch of simple heuristic controls that FastPitch enables, there's quite a lot that you can edit. For the associated code and models, have a look at our GitHub: https://github.com/DigitalPhonetics/IMS-Toucan")