Skip to content

Latest commit

 

History

History
171 lines (107 loc) · 4.44 KB

CHANGELOG.md

File metadata and controls

171 lines (107 loc) · 4.44 KB
Raw

Change Log

All notable changes to this project will be documented in this file.

The format is based on Keep a Changelog and this project adheres to Semantic Versioning.

Unreleased

[2.3.2] - 2020-11-18

Changed

  • Increase minimum version of tensorflow to v1.15.4 to fix the security vulnerability reported in https://github.com/advisories/GHSA-4g9f-63rx-5cw4 (#105)
  • Set more specific scikit-learn version requirements to avoid incompatibilities and test failures (#105)

Removed

  • Remove Python 3.5 support (#105)

[2.3.1] - 2020-04-06

Fixed

  • Fixed an issue with builds failing due to numerical issues (#103)

Changed

[2.3.0] - 2019-11-22

Changed

  • Allowed a recent version of scikit-learn (#99).

Fixed

  • Updated tests for changes in new versions of scipy, scikit-learn, and flake8 (#98).
  • Increased required version of tensorflow due to published CVEs in older versions (#98).

[2.2.0] - 2018-06-07

Added

  • Added prediction_gradient method for understanding the impact of different features in MLPs with dense inputs.
  • Added support for SELU activations with alpha dropout.
  • Added sample weights for the FMClassifier.
  • Added FMRegressor.

Fixed

  • Exposed muffnn.__version__.
  • Fixed bug in FMClassifier where it failed for predicting one example.
  • Fixed ValueError for type of target in MLPClassifier and FMClassifier (#90).

Changed

  • Updated requirements on numpy to 1.14 or higher.
  • Updated requirements on scipy to 1.0 or higher.

[2.1.0] - 2018-02-12

Added

  • Added support for the sample_weight keyword argument to the fit method of MLPClassifier and MLPRegressor (#75).

Changed

  • Switched from requiring TensorFlow 1.x to 1.4.x because 1.5.0 was causing Travis CI failures with Python 3.6 (#78).

[2.0.0] - 2018-01-17

Added

  • Added a transform_layer_index keyword and transform method to the MLPClassifier and MLPRegressor to extract features from a hidden layer (#62).

Changed

[1.2.0] - 2017-09-21

Added

  • Python 2.7 compatibility (#57).
  • Added a monitor keyword to the autoencoder (#58).
  • Added a factorization machine classifier (#50).

Changed

  • Moved to Travis instead of CircleCI (#57).
  • Upgraded to TensorFlow 1.3.X.
  • Upgraded to numpy 1.13.1.
  • Upgraded to scipy 0.19.1.
  • Upgraded to scikit-learn 0.19.0.
  • Upgraded to python 3.6.2.
  • Upgraded requirements to match python 2 properly (#59).

Fixed

  • Hid slow import of tf.contrib (#54).

[1.1.2] - 2017-05-22

Changed

  • Upgraded to TensorFlow 1.1.X.

Fixed

  • Fixed bug in grid search over solver settings.
  • Fixed bug in classes_ attribute for multilabel MLP problems.

[1.1.1] - 2017-03-27

Fixed

  • Added MANIFEST.in to fix python packaging bug.

[1.1.0] - 2017-03-27

Added

  • Add ability to set the solver and its parameters for the MLPClassifier and MLPRegressor.

Changed

  • Removed Docker build.
  • Added install_requires to setup.py.
  • Updated tests of MLP base class to silence pytest warning.

Fixed

  • Fixed score method for multilabel classification.

[1.0.0] - 2017-02-23

Changed

  • Upgraded to TensorFlow 1.0.0.

[0.2.0] - 2016-11-29

Added

  • Add an autencoder implementation.
  • Add optional monitor functionality for MLP classes, for logging, early stopping, checkpointing, etc.
  • Add top-level base class for pickling TensorFlow models.
  • Add partial fitting functionality for the MLP.
  • Add support for missing labels during multilabel classification.

Changed

  • Make sparse input more efficient in MLP classes.
  • Stop adding dropout nodes to MLP graphs if keep_prob is 1.
  • Change dropout keyword argument to keep_prob for consistency with TensorFlow.
  • Updated dependencies (notably, scikit-learn and TensorFlow).

Fixed

  • LabelEncoder in the MLPClassifier is pickled properly.
  • Fix multilabel classification, which was broken previously.

[0.1.0] - 2016-08-25

Added

  • Multilayer Perceptron Classifier and Regressor implementations.