https://github.com/cifkao/museflow
Music sequence learning toolkit
https://github.com/cifkao/museflow
deep-learning music music-information-retrieval sequence-to-sequence tensorflow
Last synced: 12 months ago
JSON representation
Music sequence learning toolkit
- Host: GitHub
- URL: https://github.com/cifkao/museflow
- Owner: cifkao
- License: bsd-3-clause
- Created: 2018-11-26T16:57:57.000Z (over 7 years ago)
- Default Branch: master
- Last Pushed: 2020-10-21T21:19:00.000Z (over 5 years ago)
- Last Synced: 2025-06-13T22:41:09.257Z (about 1 year ago)
- Topics: deep-learning, music, music-information-retrieval, sequence-to-sequence, tensorflow
- Language: Python
- Homepage:
- Size: 217 KB
- Stars: 6
- Watchers: 4
- Forks: 2
- Open Issues: 1
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# museflow
museflow is an experimental music sequence learning toolkit, built on top of TensorFlow.
The most important modules are:
- `museflow.components` – building blocks for TensorFlow models (e.g. RNN decoder)
- `museflow.encodings` – classes defining ways to encode music for use with the models
- `museflow.trainer` – a basic implementation of model loading, saving and training
- `museflow.models` – implementations of basic models (accessible via the `museflow model` command)
- `museflow.scripts` – pre- and post-processing scripts (accessible via the `museflow script` command)
To install, run:
```sh
pip install 'museflow[gpu]'
```
To install without GPU support:
```sh
pip install 'museflow[nogpu]'
```
## License
This software is distributed under the [BSD 3-Clause License](https://github.com/cifkao/museflow/blob/master/LICENSE).
Copyright 2019 Ondřej Cífka of Télécom Paris, Institut Polytechnique de Paris.
All rights reserved.