How to Build a High-Quality Text-to-Speech (TTS) System Locally with Nvidia NeMo FastPitch
After running command:
poetry install
We have to update gcc to 12 to install cython and youtokentome which are required.
conda install -c conda-forge gcc_linux-64=12
pip install cython youtokentome
and then we can run the app like
poetry run start <yourpdfile.pdf> <youroutputfile.wav>
This code includes minimal error handling. I encourage you to enhance it by addressing any issues you encounter during use.
Further if you want change pitch or speed, you can do that manipulating the spectrogram, happy coding.