network.fit(x=X_train,
y=y_train,
batch_size=batch_size,
epochs=1000,
verbose=1,
validation_data=(X_valid, y_valid),
shuffle=True,
callbacks=[early_stopping])
return network
if __name__ == '__main__':
data_dir = '/data/local/deeplearning/DeepPsychNet/abide_I_data/hdf5_data'
data_file = 'fmri_summary.hdf5'
# ['structural', 'alff', 'degree_weighted', 'eigenvector_weighted', 'falff', 'lfcd']
# metric = 'eigenvector_weighted' # 'structural' # 'autocorr' # ''entropy'
# metric = 'all' # 'structural' # 'autocorr' # ''entropy'
# metrics = ['alff', 'degree_weighted', 'eigenvector_weighted', 'falff', 'lfcd', 'all']
metrics = ['all']
save_dir = osp.join(data_dir, '5layers_holdout_all_dropout_fc')
experiment = Experiment(api_key="GVJBMG0SIOoH7zp6Lh9cW0JbB", log_code=True)
for metric in metrics:
experiment.set_filename(metric + 'do_0.2')
run(data_folder=data_dir,
data_file=data_file,
save_folder=save_dir,
metric=metric)
