Skip to content

nn~

nn~ is a bridge between Max/MSP or PureData and LibTorch (PyTorch C++ interface) for deep learning. It does not work alone and requires pretrained models.

Key Points

  • Purpose: Allows deep learning models to run inside Max/MSP or PureData.
  • Dependencies: Pretrained models in TorchScript (.ts) format.
  • Model Management:
  • Place .ts files in nn_tilde/models or any folder accessible through Max/Pd filesystem:
    • Max: Options → File Preferences
    • PureData: File → Preferences → Path
  • Since v1.6.0, some models can be downloaded via IRCAM Forum API.
  • Usage: Load a model by specifying its name as the first argument to nn~:
  • Download here or use Deken..
  • Library developed mainly by ACIDS/Ircam.
  • Report Bugs/Errors here!
  • Available for .

Articles

  • Hierarchical temporal learning for multi-instrument and orchestral audio synthesis

    by Hierarchical temporal learning for multi-instrument and orchestral audio synthesis

    Link

Videos

Similar

    Comments

    Contributors