def run_mnist(n_train=None, n_test=None, model_type="normal"):
datasize = {"n_train": n_train, "n_test": n_test}
transformer_params = {
"image_shape": 28,
"filter_shape_l1": 5, "step_shape_l1": 1, "n_l1_output": 8,
"filter_shape_l2": 5, "step_shape_l2": 1, "n_l2_output": 4,
"filter_shape_pooling": 5, "step_shape_pooling": 5
}
ensemble_params = {
"n_estimators" : 40,
"sampling_ratio" : 0.03,
"n_jobs" : -1
}
DataPath='/git/data'
batch_size=1
dataset="MNIST"
train_loader, test_loader = data_loading(DataPath,dataset,batch_size)
run(dataset, datasize, transformer_params, ensemble_params, model_type)
i += 1
else:
images_test = np.concatenate(
(images_test, np.swapaxes(images.numpy(), 3, 1)), axis=0)
y_test = np.concatenate((y_test, labels.numpy()), axis=0)
#images_test, y_test = test_set
X_test = pcanet.transform(images_test)
y_pred = classifier.predict(X_test)
return y_pred, y_test
if __name__ == "__main__":
DataPath = '/git/data'
train_loader, test_loader = data_loading(DataPath, args.dataset,
args.batchSize)
if args.gpu >= 0:
torch.cuda.set_device(args.gpu)
if args.mode == "train":
print("Training the model...")
pcanet, classifier = train(train_loader, args.dataset, args.Numlayers)
save_model(pcanet, args.dataset + "_pcanet.pkl")
save_model(classifier, args.dataset + "_classifier.pkl")
print("Model saved")
elif args.mode == "test":
pcanet = load_model(args.dataset + "_pcanet.pkl")
classifier = load_model(args.dataset + "_classifier.pkl")
def load_mnist(DataPath, dataset, batch_size):
train_loader, test_loader = data_loading(DataPath, dataset, batch_size)
return train_loader, test_loader
batch_size = 250
num_classes = 10
num_epochs = 1
learning_rate = 0.01
channels = 1
if datasets == 'CIFAR10':
input_size = 28 * 28
hidden_size = 400
batch_size = 250
num_classes = 10
num_epochs = 1
channels = 3
learning_rate = 0.01
[train_loader, test_loader] = data_loading(DataPath, datasets, batch_size)
net1 = ConvNet1(channels, hidden_size)
# values=[0.3 , 0.4 , 0.5 , 0.6 , 0.7]
# values =np.arange(0.9,0.5,-0.1)
# values=[0.06,0.08,0.1,0.12,0.14]
# start=0.65
#values =[ 0.5, 0.4, 0.3,0.2,0.1]
#start=0.8
# values =[0.3, 0.35, 0.4, 0.45,0.5,0.55]
Error = []
#for p in range(len(values)):
start = [0.95, 0.9, 0.85, 0.8, 0.75, 0.7, 0.65]
values = [0.69, 0.59, 0.49, 0.39, 0.29, 0.19, 0.09]
def load_cifar():
train, test = data_loading(DataPath, 'CIFAR10', 1)
return reshape_dataset(train, test)