LAB2

EEG classification

dataloader.py

S4b_train = np.load('./lab2/lab2_dataset/train/S4b_train.npz')
X11b_train = np.load('./lab2/lab2_dataset/train/X11b_train.npz')
S4b_test = np.load('./lab2/lab2_dataset/test/S4b_test.npz')
X11b_test = np.load('./lab2/lab2_dataset/test/X11b_test.npz')

EEGNet.py

FirstConv -> DepthwiseConv -> SeparableConv -> Classification

ELU alpha value = 0.15

Dropout probability = 0.4

main.py

Num epochs = 150

Batch size = 64

Learning rate = 0.001

Optimizer：Adam

Loss function：torch.nn.CrossEntropyLoss()

Weight decay = 0.005