.. _Build NeurEco Compression model GUI:

Build NeurEco Compression model with the GUI
#############################################

* Fill in the **Settings** tab, the build parameters are explained in the table below:

  .. figure:: ../../../../images/GUICompressionSettingsToFill.png
    :width: 700
    :alt: Settings to build a Compression model
    :align: center

    Settings to build a Compression model

* Press **Build** button

* Once the **Build** started, the **Training**, **Evaluation**, **Metrics** and **Export** panels become available. The moment the first model is saved to the checkpoint, these panels can be used as usual. 
    
.. _NeurEco Compression build parameters GUI:

Build parameters 
===========================================

.. csv-table:: Minimum Settings to build a Tabular Compression model
  :file: csv_tables/SettingsBuildTabularCompressionGUI.csv
  :header-rows: 1
  :class: longtable
  :widths: 4, 7
  :delim: ;
  :align: center

.. _NeurEco Compression advanced parameters GUI:

Advanced parameters
===========================================

.. csv-table:: Advanced Settings to build a Tabular Compression model
  :file: csv_tables/AdvancedSettingsBuildTabularCompressionGUI.csv
  :header-rows: 1
  :class: longtable
  :widths: 4, 7
  :delim: ;
  :align: center
   
.. _Normalizing the data Compression GUI:

Data normalization for Tabular Compression
===========================================

.. csv-table:: Advanced Settings for the Data normalization
  :file: csv_tables/NormalizationSettingsBuildTabularCompressionGUI.csv
  :header-rows: 1
  :class: longtable
  :widths: 4, 7
  :delim: ;
  :align: center

.. include:: ../../CommonParts/NormalizationTabularPerFeatureGUICompression.rst

.. include:: ../../CommonParts/NormalizationTabular.rst


.. _Particular cases of Build for a Tabular Compression:

Particular cases of Build for a Tabular Compression
===========================================================

.. _Select a model from a checkpoint and improve it Compression GUI:

Select a model from a checkpoint and improve it
------------------------------------------------

.. include:: ../../CommonParts/Choose model and apply final learning GUI part1.rst
It is possible to export the chosen model as it is from the checkpoint, see :std:ref:`Export NeurEco Compression model with GUI`.

.. include:: ../../CommonParts/Choose model and apply final learning GUI part2.rst


.. _Control the size of the NeurEco Compression model during build GUI:

Control the size of the NeurEco Compression model during build
------------------------------------------------------------------------ 

| It is possible to balance the number of links between the compressor and the decompressor parts of the neural network using the parameter **Compressor/decompressor size ratio** (see. :std:ref:`NeurEco Compression advanced parameters GUI`). The decompressor being in general more complex than the compressor, this ratio has to be less or equal to one and greater than zero.
| This option is particularly useful for IoT applications. It is possible to compress a sequence of measurements collected by a sensor and to reduce the quantity of transmitted data. The reduction of radioelectric transmissions reduces battery consumption and extends the autonomy of the IoT device. It can be seen in the figure below that that the quantity of transmitted data is reduced by a factor of 13. Indeed 210 measurements are replaced by 16 compression coefficients. 

This figure shows the compression/decompression models generated by a standard NeurEco Tabular compression. The number of links is well balanced between compression and decompression. However, if the user chooses to, that balance could be shifted to create a smaller compressor like shown in the figure below:

.. figure:: ../../../../images/SolutionsMbedCompression.png
  :width: 800
  :alt: SolutionsMbedCompression
  :align: center

  Controlling the size of a compression model
	
.. note::
  The size of the compressor running on a microcontroller is reduced, while the size of the decompressor is increased
   

For a detailed example of the usage of this option, see :std:ref:`Tutorial Control the size of a Compression model Python API`.