Table 2.

Model search space, as defined by a number of parameters and a corresponding range of values for each, to be explored through the Bayesian optimisation scheme of the comet.ml platform.

Parameter name	Range of values	Range type
Number of CLs	[2, 3, 4, 5, 6]	Discrete
Kernel size	[(3, 3), (4, 4), (5, 5)]	Discrete
Number of DLs	[1, 2, 3]	Discrete
Number of nodes per DL	[4, 8, 16, 32]	Discrete
Learning rate	[5e-5, 5e-3]	Continuous

CL and DL stand for convolutional and dense layer respectively.