a package for interactive machine learning
Project description
anguilla (Documentation)
anguilla
is a mapping and interactive machine learning package for digital musical instrument design in Python.
This is an early stage project. Currently, the main interface is the IML
class, which allows adding input-output pairs (IML.add
) and subsequently mapping points in the input space to outputs (IML.map
).
anguilla
is designed to be modular and hackable. An IML
object is composed of several exchangeable parts:
- an
Embedding
embeds input points into a feature space - an
NNSearch
implements nearest-neighbor search in the feature space - an
Interpolate
combines a set of output points using the distances of their corresponding input points from a neighboring query point.
anguilla server
will expose the Python API over Open Sound Control (OSC) using iipyper.
For examples and tutorials of how to use anguilla
, see our examples repo (TBC).
Install
anguilla
can be installed from PyPI:
pip install anguilla-iml
with pytorch
If you encounter an incompatibility between the PyPI versions of pytorch and faiss-cpu, try installing them both from conda (conda install -c pytorch pytorch faiss-cpu
) before pip install anguilla-iml
.
Develop
See the iil-dev repo for a recommended dev environment.
It's also possible to develop anguilla
in isolation. You will need Poetry and your Python environment manager of choice. With conda
, for example:
conda create -n anguilla-env python=3.10 poetry
conda activate anguilla-env
then:
git clone git@github.com:Intelligent-Instruments-Lab/anguilla.git
cd anguilla
poetry install
Contact
anguilla
is developed by the Intelligent Instruments Lab. Get in touch to collaborate:
◦ iil.is ◦ Facebook ◦ Instagram ◦ X (Twitter) ◦ YouTube ◦ Discord ◦ GitHub ◦ LinkedIn ◦ Email ◦
Funding
The Intelligent Instruments project (INTENT) is funded by the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (Grant agreement No. 101001848).
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
Hashes for anguilla_iml-0.3.0-py3-none-any.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | 70fff324e629dabd3ec6541b2af05c20f0c2aff57e72c412b7d475e97f23c780 |
|
MD5 | d81fb4cddcf548cdec5bb9b73b1db29a |
|
BLAKE2b-256 | 6d9d8dd0fc175cc1d7d003196dc8ee4f669ff2ba624a1ef61a2bc28b20327dad |