globalemu v1.8.1

What's Changed

• Test preprocessing by @htjb in #26

Full Changelog: v1.8.0...v1.8.1

globalemu v1.8.0

What's Changed

• Better os compatibility by @htjb in #21

Full Changelog: v1.7.3...v1.8.0

globalemu v1.7.3

What's Changed

• Updating the citation on the readme by @htjb in #20

Full Changelog: v1.7.2...v1.7.3

globalemu v1.7.2

What's Changed

• Flush prints during network training by @ThomasGesseyJones in #14
• Better early stopping by @htjb in #15
• Version increment and Release workflows by @ThomasGesseyJones in #16
• Check Tag and Release by @ThomasGesseyJones in #17
• Fix? Tag and Release by @ThomasGesseyJones in #18

Full Changelog: v1.6.0...v1.7.2

globalemu v1.6.0

Includes a few bug updates and updated early stopping based on the test loss.

globalemu v1.5.0 release

Stable release:

• Added code so that users can train with their own custom loss functions built with tensforlow's keras backend.

globalemu v1.4.0 release

Stable release:

• Added greater freedoms for the user to the preprocessing of the data via the new kwargs AFB, resampling and std_division.
• Modified testing and documentation accordingly.

globalemu v1.3.0 release

Stable release:

• Redesigned GUI so that it relies on configuration files associated with trained instances of the emulator. This provides a new level of flexibility in that a GUI configuration file can be produced for any trained model and that trained model can subsequently be explored with the GUI.
• Added relevant docs and testing for the new functionality.

globalemu v1.2.0 release

Stable release:

• Added functionality such that multiple sets of parameters can be emulated with one call to the emulator as opposed to via a set of calls in a for loop.
• Added a function to plot the mean, 95th percentile and worst emulations from a set of input parameters and corresponding signals so that users can easily assess the quality of the training.
• Restructuring of cmSim.py.
• Additional tutorials and tests for new functionality.

globalemu v1.1.1 release

Stable release:

• Updated GUI to be more compact
• Corrected GUI slider ranges
• Added more detail to README and docs about the released models
• Corrected spelling mistake in kwarg error reporting