Changes in version 1.3.2 (2020-01-16)                  

    o   one step forward, one step back!
	MAPE cost function is back and validated
	RMSE cost function removed (derivative
	challenge + no added value vs MSE imo)

    o   bug fix: 'chkgradevery' is now available

                 Changes in version 1.3.1 (2020-01-09)                  

    o   direct ‘overfitting’  &  ‘regularization’
	modes for auto_runtype

    o   batch normalization set to 0 by default
	(not good for regression targets)

    o   bug fix: now exits if no gain when no cros-
	s validation sample

    o   bug fix: testgainunder now kept as is

    o   MAPE cost function removed (probably an er-
	ror to be fixed in derivative: todo list)

                 Changes in version 1.3.0 (2019-12-15)                  

    o   rmse & mape cost function added for regres-
	sion

                        Changes in version 1.2.9                        

    o   documentation update (typo correction and
	more detailed explanations concerning
	modexec and autopar)
	thanks to Alexey

                 Changes in version 1.2.8 (2019-03-16)                  

    o   time limit for sub modelizations in automl_
	train function, to avoid waiting too long
	for a specific particle to finish its mode-
	lization

                 Changes in version 1.2.7 (2019-01-27)                  

    o   lightening models produced by removing
	unnecessary data (Z, A layers, etc...)

                 Changes in version 1.2.6 (2018-12-08)                  

    o   just a proper Authors@R field 4 CRAN :-)

                        Changes in version 1.2.5                        

    o   automl_train bug fix when particle produces
	no exploitable model (continued work)

                        Changes in version 1.2.4                        

    o   automl_train bug fix when particle produces
	no exploitable model

                        Changes in version 1.2.3                        

    o   dropout bug fix (now random at each mini
	batch and reproductible)

    o   bug fix: continue training on auto trained model
	with manual function

                        Changes in version 1.2.2                        

    o   pkgdown site with vignettes (Rmd file added)

                        Changes in version 1.2.1                        

    o   pkgdown site (nb: Rnw vignettes are not included)

                 Changes in version 1.2.0 (2018-11-23)                  

    o   New param 'mdlref' for automl_train_manual:
	to start training from saved model (shape,
	weights...) for fine tuning

    o   New param 'mdlref' for automl_train:
	to start training with loaded hpar and autopar
	(not the model)

    o   New 'auto_runtype' autopar for automl_train
	to run automatically the 2 steps below;
	1: overfitting, goal: performance
	2: regularization, goal: generalization

    o   dropout bug fix

    o   lambda bug fix

                        Changes in version 1.1.0                        

    o   same train/test sampling for each particle with automl_train

    o   stick to a format in variables naming

                        Changes in version 1.0.9                        

    o   testcvsize = 0 bug fix

                 Changes in version 1.0.8 (2018-11-08)                  

    o   New automatic hyperparameters adjustments below:
	'auto_psovelocitymaxratio' autopar PSO velocity max ratio
	'auto_layer' autopar layer shape (layers number, nodes
	number per layer, activation types and dropout ratios)

    o   'auto_lambda' bug fix

                        Changes in version 1.0.7                        

    o   vignette completion

                 Changes in version 1.0.6 (2018-10-21)                  

    o   vignette howto_automl: why, how and basic howto

    o   'auto_lambda' autopar regularization hyperparameter

    o   seed bug fix for reproductibility

    o   display enhancement in NN structure display

                 Changes in version 1.0.5 (2018-09-13)                  

    o   first public release on CRAN

    o   There's so much to do; transfert learning, CNN, RNN ...
	feel free to join