🎵 Generate Music in Minutes: Introducing ComfyUI_DiffRhythm v1.0.0!

ComfyUI_DiffRhythm: Fast and Easy End-to-End Full-Length Song Generation

GitHub Repository: https://github.com/billwuhao/ComfyUI_DiffRhythm

📣 Update [2025-03-13] ⚒️: Version v1.0.0 released. All parameters are optional, and music can be generated randomly without any input.

Model Download

Models will be automatically downloaded to the folder:
ComfyUI\models\TTS\DiffRhythm

Folder Structure:

• Manual Download Links:

  • DiffRhythm-base: cfm_model.pt

  • DiffRhythm-vae: vae_model.pt

  • MuQ-MuLan-large: OpenMuQ/MuQ-MuLan-large

  • MuQ-large-msd-iter: OpenMuQ/MuQ-large-msd-iter

  • xlm-roberta-base: FacebookAI/xlm-roberta-base

Environment Configuration

For Windows Systems (other systems like Linux and Mac are untested but should be supported):

1. Download and Install the Latest Version of espeak-ng:
   Download from: espeak-ng v1.52.0

2. Add Environment Variable:
   Add the environment variable PHONEMIZER_ESPEAK_LIBRARY to your system. Set its value to the path of the libespeak-ng.dll file in your espeak-ng installation directory. For example:

   复制

   C:\Program Files\eSpeak NG\libespeak-ng.dll

Acknowledgments

• DiffRhythm: https://github.com/ASLP-lab/DiffRhythm

With ComfyUI_DiffRhythm, you can now generate full-length songs quickly and easily using ComfyUI nodes. Dive in and start creating your own music today! 🚀