Deep learning extension package for sktime, a scikit-learn compatible toolbox for learning with time series data
The package is under active development. Currently, classification models based off the the networks in dl-4-tsc have been implemented, as well as an example of a tuned network for future development.
This package uses the base sktime as a dependency. Follow the original instructions to install this.
For the deep-learning part of sktime-dl, you need:
and a compatible backend for Keras, one of
If you want to run the networks on a GPU, CUDNN is also required to be able to utilise your GPU.
For windows users, we recommend following this (unaffiliated) guide.
For linux users, all of these points should hopefully be relatively straight forward via simple pip-commands and conversions from the previous link.
For mac users, I am unfortunately unsure of the best processes for installing these. If you have links to a tested and up-to-date guide, let us know (@James-Large).
A repository for off-the-shelf networks
The aim is to define Keras networks able to be directly used within sktime and its pipelining and strategy tools, and by extension scikit-learn, for use in applications and research. Overtime, we wish to interface or reimplement networks from the literature in the context of time series analysis.
Currently, we interface with a number of networks for time series classification in particular.
This toolset currently serves as an interface to dl-4-tsc, and implements the following network archtiectures:
- Time convolutional neural network (CNN)
- Encoder (Encoder)
- Fully convolutional neural network (FCNN)
- Multi channel deep convolutional neural network (MCDCNN)
- Multi-scale convolutional neural network (MCNN)
- Multi layer perceptron (MLP)
- Residual network (resnet)
- Time Le-Net (tlenet)
- Time warping invariant echo state network (twiesn)
Documentation for sktime-dl shall be produced in due course.
Former and current active contributors are as follows.
James Large (@James-Large), Aaron Bostrom (@ABostrom), Hassan Ismail Fawaz (@hfawaz), Markus Löning (@mloning)
Project management: Jason Lines (@jasonlines), Franz Király (@fkiraly)
Design: Anthony Bagnall(@TonyBagnall), Sajaysurya Ganesh (@sajaysurya), Jason Lines (@jasonlines), Viktor Kazakov (@viktorkaz), Franz Király (@fkiraly), Markus Löning (@mloning)
Coding: Sajaysurya Ganesh (@sajaysurya), Bagnall(@TonyBagnall), Jason Lines (@jasonlines), George Oastler (@goastler), Viktor Kazakov (@viktorkaz), Markus Löning (@mloning)
We are actively looking for contributors. Please contact @fkiraly or @jasonlines for volunteering or information on paid opportunities, or simply raise an issue in the tracker.
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
|Filename, size||File type||Python version||Upload date||Hashes|
|Filename, size sktime_dl-0.1.0-py3-none-any.whl (33.4 kB)||File type Wheel||Python version py3||Upload date||Hashes View|
|Filename, size sktime-dl-0.1.0.tar.gz (22.4 kB)||File type Source||Python version None||Upload date||Hashes View|
Hashes for sktime_dl-0.1.0-py3-none-any.whl