Skip to main content

Table 1 The hyper-parameters of the model and their corresponding values

From: MCGNet+: an improved motor imagery classification based on cosine similarity

Hyperparameter Value
Learning rate 9.6e–4
Learning rate decay 0
Dropout rate 0.5
Optimizer Adam
L1, L2 regularization 0.002, 0.001
Training epochs 500
Batchsize 32
Chebyshev polynomial 2