Table 2. Hyperparameters of the multitask 1D-convolutional neural network–based architecture.

Layer	Type	Hyperparameters
Input	Input	Features size
Conv1D	Convolution	Input=1, output=20, K=8, stride=1, padding
ReLu	Activation function	ReLu
Pooling	Pooling	Stride=2, max pooling
Conv	Convolution layer	Input=20, output=40, K=8, stride=1, padding
Relu	Activation function	ReLu
Pooling	Pooling	Stride=2, max pooling
Conv	Convolution layer	Input=40, output=60, K=8, stride=1, padding
Relu	Activation function	ReLu
Pooling	Pooling	Stride=2, max pooling
Fully connected 1	Fully connected layer	Input=360, output=100
Fully connected 2	Fully connected layer	Input=100, output=300
Linear 1	Input: 100, output: 2	Output=2 affect classes, activation function:linear
Linear 2	Input: 300, output: N	Output=N identity classes, activation function:linear