Home
Spleeter is the Deezer source separation library. It comes with pretrained models, is written in Python and uses Tensorflow (only v1 is supported for now). It makes it easy to train source separation model (assuming you have a dataset of isolated sources), and provides already trained state of the art model for performing various flavour of separation.
The separation process can be performed both on GPU or CPU. It is very fast when run on GPU: for instance, the whole musDB test set (about 3h27m of audio) is separated in less than 90s on a single Nvidia GeForce GTX 1080 (audio loading and writing is done on a 32 core Intel(R) Xeon(R) Gold 6134 CPU @ 3.20GHz).
This documentation provides:
- Installation instructions
- Usage manual to get you started
- A deeper view of the Python API and the pretrained Models packaged with Spleeter
