Skip to content

cryptobro-man/heartmula_comfyui

 
 

Repository files navigation

HeartMuLa_ComfyUI

ComfyUI Custom Node for HeartMuLa AI Music Generation and Transcript Text

HeartMuLa official GITHUB https://github.com/HeartMuLa/heartlib

How To Use this In Basic: https://youtu.be/F9LFAeUbBIs

image e48e0c00-70eb-4ee3-b578-784ff92b1d0c

Message

Let's make this project the true open source, anyone who is interested in making improvements feel free to let us know in Discussions : the https://github.com/benjiyaya/HeartMuLa_ComfyUI/discussions

We are not providing service here, we are group of hobbiests, developers who want to make something here. So don't take it for granted.


Update:


2026-01-21 (3)

  • Integrates native progress bars, making it easy to implement real-time progress tracking in the user interface.
  • Precise Temperature: Refined temperature step to 0.01 for more granular control over generation.
  • Intuitive Audio Length: Renamed parameter to max_audio_length_seconds (Default: 240s) for better usability.
  • Keep Model Loaded(Memory Settings): True: Keep model in VRAM for instant subsequent generations. False: Unload model after each task to free up memory.
  • Offload Mode: Auto: Standard memory release for balanced performance. Aggressive: Full VRAM wipe + Garbage Collection.

2026-01-21 (2)

  • Lazy Load Optimization , now able to load with 12GB VRAM.
  • Path Configuration ,support custom model folder path in in the "extra_model_paths.yaml", Not limited by default ComfyUI/Models/ folder path only.

2026-01-21

  • MEMORY CLEANUP and Pipeline changed for BF16 - Optimized for 16GB. dtype I don't recommand under bf16 for this model, audio quality will degrade too much.

Installation


Step 1

Go to ComfyUI\custom_nodes Command prompt:

git clone https://github.com/benjiyaya/HeartMuLa_ComfyUI

Step 2

cd /HeartMuLa_ComfyUI

Step 3

pip install -r requirements.txt

If no module name error pop up. some libraries might need to install Individually (For Windows users you need use Command Prompt as Administrator)

do this :

pip install soundfile

pip install torchtune

pip install torchao


For File structure


image

Download model files


Go to ComfyUI/models

Use HuggingFace Cli download model weights.

type :

hf download HeartMuLa/HeartMuLaGen --local-dir ./HeartMuLa

hf download HeartMuLa/HeartMuLa-oss-3B --local-dir ./HeartMuLa/HeartMuLa-oss-3B

hf download HeartMuLa/HeartCodec-oss --local-dir ./HeartMuLa/HeartCodec-oss

hf download HeartMuLa/HeartTranscriptor-oss --local-dir ./HeartMuLa/HeartTranscriptor-oss


For Model File structure


image

Model Sources

Github Repo: https://github.com/HeartMuLa/heartlib

Paper: https://arxiv.org/abs/2601.10547

Demo: https://heartmula.github.io/

HeartMuLa-oss-3B: https://huggingface.co/HeartMuLa/HeartMuLa-oss-3B

HeartCodec-oss: https://huggingface.co/HeartMuLa/HeartCodec-oss

HeartTranscriptor-oss: https://huggingface.co/HeartMuLa/HeartTranscriptor-oss

Credits

HeartMuLa: https://huggingface.co/HeartMuLa/HeartMuLa-oss-3B

About

ComfyUI Custom Node for HeartMuLa AI Music Generation and Transcript Text

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors

Languages

  • Python 100.0%