This is a repo of scalp EEG classification.
data description this data resample from ~300k raw EEG segments by keeping the type ratios.
1.type_0 refer to background
2.type_1 refer to IED
3.type_2 refer to artifacts
parameters description
1.train params, include number of iteration, maxing learning rate
2.architecture elecments params, inlude layers, filter sizes, dropout...
before running
1.pip install pytorch
2.pip install fastai
3.pip install scikit,numpy
4.pip install tasi