A Fortran-Keras Deep Learning Bridge for Scientific Computing

Table 2

Spearman correlation of corresponding hyperparameter with online performance and associated value.

	Correlation	-value
BatchNorm	0.0859	3.7896e−01
Dropout	0.1919	4.7591e−02
Leaky ReLU	0.0055	9.5465e−01
Learning rate	−0.2087	3.0923e−02
Dense nodes	0.1427	1.4249e−01
Layers	0.0410	6.7491e−01
Optimizer	−0.6998	5.0177e−17
Parameters	0.1528	1.1609e−01