Skip to content

Latest commit

 

History

History
68 lines (60 loc) · 10.9 KB

params.md

File metadata and controls

68 lines (60 loc) · 10.9 KB

Parameters

Model Configuration Parameters Comparison

MEEG

Parameter AT-DGNN LGGNet EEGNet DeepConvNet ShallowConvNet EEG-TCNet TSception TCNet-Fusion ATCNet DGCNN
segment 4 4 4 4 4 4 4 4 4 4
overlap 0 0 0 0 0 0 0 0 0 0
sampling-rate 1000 1000 1000 1000 1000 1000 1000 1000 1000 1000
target-rate 200 200 200 200 200 200 200 200 200 200
trial-duration 59 59 59 59 59 59 59 59 59 59
input-shape (1,32,800) (1,32,800) (1,32,800) (1,32,800) (1,32,800) (1,32,800) (1,32,800) (1,32,800) (1,32,800) (1,32,800)
channels 32 32 32 32 32 32 32 32 32 32
fold 10 10 10 10 10 10 10 10 10 10
max-epoch 200 200 200 400 400 200 200 200 400 400
patient 20 20 20 40 20 20 20 20 40 40
patient-cmb 8 8 8 10 8 8 8 8 20 20
max-epoch-cmb 20 20 20 40 20 20 20 20 40 40
batch-size 64 64 64 64 64 64 64 64 64 64
learning-rate 1e-03 1e-03 1e-03 1e-05 1e-05 1e-03 1e-05 1e-05 1e-05 1e-04
training-rate 0.8 0.8 0.8 0.8 0.8 0.8 0.8 0.8 0.8 0.8
weight-decay 0.001 0.001 0.001 0.001 0.001 0.001 0.001 0.001 0.001 0.001
step-size 5 5 5 5 5 5 5 5 5 5
dropout 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5
LS Label smoothing Label smoothing Label smoothing Label smoothing Label smoothing Label smoothing Label smoothing Label smoothing Label smoothing Label smoothing
LS-rate 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1
pool 16 16 16 16 16 16 16 16 16 16
pool-step-rate 0.25 0.25 0.25 0.25 0.25 0.25 0.25 0.25 0.25 0.25
T 64 64 64 64 64 64 64 64 64 64
graph-type BL BL BL BL BL BL BL BL
hidden 32 32 32 32 32 32 32 32 32 32

If you want to run the code on MEEG dataset, please set the parameters as the table above. You can also refer to the run to run the code.

DEAP

Parameter AT-DGNN LGGNet EEGNet DeepConvNet ShallowConvNet EEG-TCNet TSception TCNet-Fusion ATCNet DGCNN
segment 4 4 4 4 4 4 4 4 4 4
overlap 0 0 0 0 0 0 0 0 0 0
sampling-rate 128 128 128 128 128 128 128 128 128 128
target-rate 128 128 128 128 128 128 128 128 128 128
trial-duration 63 63 63 63 63 63 63 63 63 63
input-shape (1,32,512) (1,32,512) (1,32,512) (1,32,512) (1,32,512) (1,32,512) (1,32,512) (1,32,512) (1,32,512) (1,32,512)
channels 32 32 32 32 32 32 32 32 32 32
fold 10 10 10 10 10 10 10 10 10 10
max-epoch 200 200 200 400 400 200 200 200 400 400
patient 20 20 20 40 20 20 20 20 40 40
patient-cmb 8 8 8 10 8 8 8 8 20 20
max-epoch-cmb 20 20 20 40 20 20 20 20 40 40
batch-size 64 64 64 64 64 64 64 64 64 64
learning-rate 1e-03 1e-03 1e-03 1e-05 1e-05 1e-03 1e-05 1e-05 1e-05 1e-04
training-rate 0.8 0.8 0.8 0.8 0.8 0.8 0.8 0.8 0.8 0.8
weight-decay 0.001 0.001 0.001 0.001 0.001 0.001 0.001 0.001 0.001 0.001
step-size 5 5 5 5 5 5 5 5 5 5
dropout 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5
LS Label smoothing Label smoothing Label smoothing Label smoothing Label smoothing Label smoothing Label smoothing Label smoothing Label smoothing Label smoothing
LS-rate 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1
pool 16 16 16 16 16 16 16 16 16 16
pool-step-rate 0.25 0.25 0.25 0.25 0.25 0.25 0.25 0.25 0.25 0.25
T 64 64 64 64 64 64 64 64 64 64
graph-type BL BL BL BL BL BL BL BL
hidden 32 32 32 32 32 32 32 32 32 32

If you want to run the code on DEAP dataset, please set the parameters as the table above. You can also refer to the run to run the code.