seed = 42
sparse = False
Nclass = 10
k = 10000
nn_cache = "../Data/hw2-data/MNIST_nn_cache.npz"
if not os.path.exists(nn_cache):
RuntimeError("Transformation matrix v doesn't exist!")
else:
print("Reading from cache...")
res = np.load(nn_cache)
v = res["v"]
# Load in MNIST training data then transform it
print("Loading MNIST Training data...")
X_train, y_train = mu.load_mnist(dataset='training')
X_train = ru.naive_nn_layer(X_train, k=k, v=v)
y_train_true = np.asarray(y_train[:, None] == np.arange(max(y_train)+1),dtype=int).squeeze()
print("Estimated best lambda: %.3lf" % val.estimate_lambda(X_train, scale=1.0e-2))
# Load in MNIST training data then transform it
print("Loading MNIST Testing data...")
X_test, y_test = mu.load_mnist(dataset='testing')
X_test = ru.naive_nn_layer(X_test, k=k, v=v)
y_test_true = np.asarray(y_test[:, None] == np.arange(max(y_test)+1),dtype=int).squeeze()
#######################################################
#
# Run minibatch with bestfit?
#
#######################################################
import matplotlib.pyplot as plt
import matplotlib as mpl
#Typical plot parameters that make for pretty plots
mpl.rcParams['figure.figsize'] = (9,8)
mpl.rcParams['font.size'] = 20.0
mpl.rc('font',**{'family':'serif','serif':['Computer Modern']})
mpl.rc('text', usetex=True)
# Flags to control functionality
show_plots = True
save_plots = True
use_one_digit = False
# Load in MNIST data
print("Loading MNIST Training data...")
X_train, y_train = mu.load_mnist(dataset='training')
# Training using only one digit?
if use_one_digit:
mask = y_train.squeeze() == 5
X_train = X_train[mask]
print(X_train.shape)
y_train = y_train[mask]
# Init PCA object
# Solve for all principal components but do calculations using only 50
# Can reset l later if need be as all principal components are retained
PCA = pca.PCA(l=50, center=True)
# Fit model
print("Fitting PCA model...")
# Test mnist data importing functions
import sys
sys.path.append("..")
import DML.data_processing.mnist_utils as mu
# Test data loading, make sure shape is correct
images, labels = mu.load_mnist(dataset='training')
print(images.shape, labels.shape)
# Test mnist data importing functions
import sys
sys.path.append("..")
import DML.data_processing.mnist_utils as mu
# Test data loading, make sure shape is correct
images, labels = mu.load_mnist(dataset='training')
print(images.shape,labels.shape)
nout = 30000
nclass = 10
nodes = 500
seed = 42
max_iters = 50
# Seed RNG
np.random.seed(seed=seed)
# Flags to control functionality
show_plots = True
save_plots = True
# Load in MNIST data, process it for multiclass classification
print("Loading MNIST data...")
X_train, y_train = mu.load_mnist(dataset='training')
y_train_true = np.asarray(y_train[:, None] == np.arange(max(y_train) + 1),
dtype=int).squeeze()
X_test, y_test = mu.load_mnist(dataset='testing')
y_test_true = np.asarray(y_test[:, None] == np.arange(max(y_test) + 1),
dtype=int).squeeze()
# Solve for all principal components but do calculations using only 50
# Can reset l later if need be as all principal components are retained
print("Performing PCA with k = %d components..." % k)
PCA = pca.PCA(l=k, center=False)
PCA.fit(X_train)
X_train = PCA.transform(X_train)
X_test = PCA.transform(X_test)
print("Training neural network...")
import matplotlib.pyplot as plt
import matplotlib as mpl
# Typical plot parameters that make for pretty plots
mpl.rcParams["figure.figsize"] = (9, 8)
mpl.rcParams["font.size"] = 20.0
mpl.rc("font", **{"family": "serif", "serif": ["Computer Modern"]})
mpl.rc("text", usetex=True)
# Flags to control functionality
show_plots = True
save_plots = True
# Load in MNIST data
print("Loading MNIST data...")
X_train, y_train = mu.load_mnist(dataset="training")
y_train_true = np.asarray(y_train[:, None] == np.arange(max(y_train) + 1), dtype=int).squeeze()
# print("Loading MNIST Testing data...")
X_test, y_test = mu.load_mnist(dataset="testing")
y_test_true = np.asarray(y_test[:, None] == np.arange(max(y_test) + 1), dtype=int).squeeze()
# Init PCA object
# Solve for all principal components but do calculations using only 50
# Can reset l later if need be as all principal components are retained
PCA = pca.PCA(l=50, center=True)
print("Training the model...")
PCA.fit(X_train)
X_train = PCA.transform(X_train)
X_test = PCA.transform(X_test)
