Release 0.2.4

What's Changed

• TF 2.5 compatibility by @vdutor in #48
• Place upper bound on TensorFlow (Probability) dependencies by @ltiao in #52
• Make RFF weights explicitly not trainable by @ltiao in #51
• Refactoring basis functions by @ltiao in #53
• Added support for alternative Fourier feature map by @ltiao in #54
• Quadrature Fourier features by @ltiao in #56
• Orthogonal Random Features by @ltiao in #57

New Contributors

• @ltiao made their first contribution in #52

Full Changelog: v0.2.3...v0.2.4

Release 0.2.3

Release 0.2.3

Bugfixes

Fix PyPi upload Github Action

Thanks to our Contributors

This release contains contributions from (alphabetical order)

@sebastianober, @vdutor

Release 0.2.2

Release 0.2.2

Bugfixes

• Fix PyPi upload Github Action (#46)

Thanks to our Contributors

This release contains contributions from (alphabetical order)

@sebastianober, @vdutor

Release 0.2.1

Release 0.2.1

Bugfixes

• Fix PyPi upload Github Action (#45)

Thanks to our Contributors

This release contains contributions from (alphabetical order)

@sebastianober, @vdutor

Release 0.2.0

Release 0.2.0

Improvements

• Allow for whitening in sampling methods (#26)
• Add warning for default Identity mean function in GPLayer (#42)
• Allow the user to specify which layers to train with NaturalGradient (#43)

Documentation

• Update README with link to GPflux paper (#22)
• Clean notebook deep_gp_samples notebook (#23)
• Fix header in efficient_sampling notebook (#24)
• New notebook on conditional density estimation with GPflux (#40)
• Improve plotting in gpflux_with_keras_layers (#41)

Thanks to our Contributors

This release contains contributions from (alphabetical order)

@johnamcleod, @sebastianober, @st--, @vdutor