MoonInTheRiver / DiffSinger

DiffSinger: Singing Voice Synthesis via Shallow Diffusion Mechanism (SVS & TTS); AAAI 2022; Official code
MIT License
4.26k stars 713 forks source link

Migration to new code structures #70

Closed cyclekiller closed 2 years ago

cyclekiller commented 2 years ago
  1. combine all config files into configs/
  2. move tts-specific files out (partially done), delete ljspeech processed data as it takes too much space
  3. combine tasks/singing, vocoders/ and usr/ into src/, as they are related to the core pipeline
  4. move base_task.py, base_binarizer.py out into basics/, as they are the base classes which define the interface for other classes
  5. some minor changes such as moving inference/svs/ to inference/, moving the entry point run.py out, etc.
  6. change references in all the files, including docs The result runs sucessfully on my machine.

For users with Colab or Huggingface, simply replacing .../run.py with run.py and usr/config/... with config/... should be enough.