CNN-Based Personal Identification System Using Resting State Electroencephalography

Table 5

Model loading time (T_model) and averaged execution time for batch testing (T_batch) for 1D-Convolutional LSTM and the proposed approach.

Model	Channels	Tmodel (s)	Tbatch (s)
1D-Convolutional LSTM [27] (TensorFlow)	16	17.852	0.065
	32	17.965	0.065
	64	18.477	0.071
Proposed (PyTorch)	14	1.106	0.002
	32	1.102	0.002
	64	1.125	0.002