---
license: mit
library_name: pytorch
tags:
- variational-autoencoder
- audio
- music-generation
- encodec
---

# VAE Drum Machine — data & checkpoints

Trained checkpoints and processed data for a variational autoencoder that generates and
interpolates drum one-shots in Meta EnCodec token space.

Code and full write-up: https://github.com/kevin-bretz/VAEDrumMachine

**Checkpoints** (`checkpoints/`)
- `model.pt` — trained VAE.
- `vae_encodec_checkpoint_epoch_660.pt`, `vae_encodec_checkpoint_epoch_840.pt` — intermediate training checkpoints.

**Data** (`data/`)
- `encodec_cache.zip` — EnCodec token cache.
- `Colab_Data.zip` — processed training data used on Colab.