Skip to content

StemGen

Generate music that responds to what it hears

Visit official site

What StemGen is

StemGen is described as an end-to-end music generation model trained to listen to a musical context and respond appropriately. The model is built on a non-autoregressive language model style architecture similar to SoundStorm and VampNet, and the page shows several example outputs.

What you can do with it

  • Review the model and dataset material
  • Listen to test set examples
  • Work through iterative generation examples
  • Start generation from drums or from chords

Application scenarios

  • Build a website that shows the model output examples
  • Create a real-time interactive music generation mini app
  • Develop a music generation desktop client around the model

Model notes

  • Generation responds to the musical context it listens to
  • The architecture is non-autoregressive like SoundStorm and VampNet
  • Deep iterative layering is listed among the examples

Pros & cons

✓ What we like

  • Generation can start from drums or from chords
  • The page lists test set and iterative generation examples

! What to watch out for

  • The architecture follows a non-autoregressive language model style
  • Generation is conditioned on the musical context it hears

FAQ

What architecture does it use?

The material describes a non-autoregressive language model style architecture similar to SoundStorm and VampNet.

What examples are on the page?

The feature list names test set examples, iterative generation examples, starting from drums, starting from chords and deep iterative layering.

Last reviewed: 2026-11-15

More AI music creation tools

View all →

How we review