def prepro(X_train, X_val, X_test):
mean = np.mean(X_train)
return X_train - mean, X_val - mean, X_test - mean
if __name__ == '__main__':
if len(sys.argv) > 1:
net_type = sys.argv[1]
valid_nets = ('ff', 'cnn')
if net_type not in valid_nets:
raise Exception('Valid network type are {}'.format(valid_nets))
else:
net_type = 'ff'
mnist = input_data.read_data_sets('data/MNIST_data/', one_hot=False)
X_train, y_train = mnist.train.images, mnist.train.labels
X_val, y_val = mnist.validation.images, mnist.validation.labels
X_test, y_test = mnist.test.images, mnist.test.labels
M, D, C = X_train.shape[0], X_train.shape[1], y_train.max() + 1
X_train, X_val, X_test = prepro(X_train, X_val, X_test)
if net_type == 'cnn':
img_shape = (1, 28, 28)
X_train = X_train.reshape(-1, *img_shape)
X_val = X_val.reshape(-1, *img_shape)
X_test = X_test.reshape(-1, *img_shape)
solvers = dict(sgd=sgd,
n_experiment = 1
reg = 1e-5
print_after = 50
p_dropout = .8
loss = 'cross_ent'
nonlin = 'relu'
solver = 'sgd'
solver3 = 'sgd3'
#worker_num =10
filename1 = './1.txt'
filename2 = './2.txt'
f1 = open(filename1,'w')
f2 = open(filename2,'w')
mnist = input_data.read_data_sets('MNIST_Data/',one_hot = True)
def prepro(X_train, X_val, X_test):
mean = np.mean(X_train)
return X_train - mean, X_val - mean, X_test - mean
if __name__ == '__main__':
if len(sys.argv) > 1:
net_type = sys.argv[1]
valid_nets = ('ff', 'cnn')
if net_type not in valid_nets:
raise Exception('Valid network type are {}'.format(valid_nets))
else:
net_type = 'cnn'
def prepro(X_train, X_val, X_test):
mean = np.mean(X_train)
return X_train - mean, X_val - mean, X_test - mean
if __name__ == '__main__':
if len(sys.argv) > 1:
net_type = sys.argv[1]
valid_nets = ('ff', 'cnn')
if net_type not in valid_nets:
raise Exception('Valid network type are {}'.format(valid_nets))
else:
net_type = 'ff'
mnist = input_data.read_data_sets('data/MNIST_data/', one_hot=False)
X_train, y_train = mnist.train.images, mnist.train.labels
X_val, y_val = mnist.validation.images, mnist.validation.labels
X_test, y_test = mnist.test.images, mnist.test.labels
M, D, C = X_train.shape[0], X_train.shape[1], y_train.max() + 1
X_train, X_val, X_test = prepro(X_train, X_val, X_test)
if net_type == 'cnn':
img_shape = (1, 28, 28)
X_train = X_train.reshape(-1, *img_shape)
X_val = X_val.reshape(-1, *img_shape)
X_test = X_test.reshape(-1, *img_shape)
solvers = dict(