import pickle from pymri.dataset import load_nnadl_dataset mode = 0 if mode == 0: path = '/home/jesmasta/amu/master/nifti/bold/' training_data, validation_data, test_data = load_nnadl_dataset( path, # (('ExeCtrl_0', 'ExeCtrl_5'), ('ExeTool_0', 'ExeTool_5')), (('ExeTool_0', 'ExeTool_5'), ('ExeCtrl_0', 'ExeCtrl_5')), k_features=784, normalize=True, scale_0_1=False, sizes=(0.5, 0.25, 0.25)) pickle.dump([training_data, validation_data, test_data], open("/tmp/save.p", "wb")) elif mode == 1: training_data, validation_data, test_data = pickle.load( open("/tmp/save.p", "rb")) def perform(input_features=784): # from pymri.model import fnn # net = fnn.Network([input_features, 46, 2]) # net.SGD(training_data, 100, 11, 2.961, test_data=test_data) import fnn2
import pickle from pymri.dataset import load_nnadl_dataset mode = 0 if mode == 0: path = '/home/jesmasta/amu/master/nifti/bold/' training_data, validation_data, test_data = load_nnadl_dataset( path, # (('ExeCtrl_0', 'ExeCtrl_5'), ('ExeTool_0', 'ExeTool_5')), (('ExeTool_0', 'ExeTool_5'), ('ExeCtrl_0', 'ExeCtrl_5')), k_features = 784, normalize=True, scale_0_1=False, sizes=(0.5, 0.25, 0.25) ) pickle.dump( [training_data, validation_data, test_data], open("/tmp/save.p", "wb") ) elif mode == 1: training_data, validation_data, test_data = pickle.load( open("/tmp/save.p", "rb") ) def perform(input_features=784):
import pickle from pymri.dataset import load_nnadl_dataset mode = 0 if mode == 0: path = "/home/jesmasta/amu/master/nifti/bold/" training_data, validation_data, test_data = load_nnadl_dataset( path, # (('ExeCtrl_0', 'ExeCtrl_5'), ('ExeTool_0', 'ExeTool_5')), (("ExeTool_0", "ExeTool_5"), ("ExeCtrl_0", "ExeCtrl_5")), sizes=(0.75, 0.25), ) pickle.dump([training_data, validation_data, test_data], open("/tmp/save.p", "wb")) elif mode == 1: training_data, validation_data, test_data = pickle.load(open("/tmp/save.p", "rb")) def perform(input_features=784): from pymri.model import fnn # net = fnn.Network([784, 30, 2]) # net.SGD(training_data, 100, 10, 3.0, test_data=test_data) net = fnn.Network([input_features, 46, 2]) net.SGD(training_data, 100, 11, 2.961, test_data=test_data) return net
import pickle from pymri.dataset import load_nnadl_dataset mode = 0 if mode == 0: path = '/home/jesmasta/amu/master/nifti/bold/' training_data, validation_data, test_data = load_nnadl_dataset( path, # (('ExeCtrl_0', 'ExeCtrl_5'), ('ExeTool_0', 'ExeTool_5')), (('ExeTool_0', 'ExeTool_5'), ('ExeCtrl_0', 'ExeCtrl_5')), k_features = 784, sizes=(0.75, 0.25) ) pickle.dump( [training_data, validation_data, test_data], open("/tmp/save.p", "wb") ) elif mode == 1: training_data, validation_data, test_data = pickle.load( open("/tmp/save.p", "rb") ) def perform(input_features=784): from pymri.model import fnn # net = fnn.Network([784, 30, 2])