def main():
Xtrain, Ytrain, Xtest, Ytest = getKaggleMNIST()
# dnn = DNN([1000, 750, 500])
# dnn.fit(Xtrain, Ytrain, Xtest, Ytest, epochs=3)
# vs
dnn = DNN([1000, 750, 500])
dnn.fit(Xtrain, Ytrain, Xtest, Ytest, pretrain=False, epochs=10)
def main(loadfile=None, savefile=None):
Xtrain, Ytrain, Xtest, Ytest = getKaggleMNIST()
if loadfile:
dbn = DBN.load(loadfile)
else:
dbn = DBN([1000, 750, 500, 10]) # AutoEncoder is default
# dbn = DBN([1000, 750, 500, 10], UnsupervisedModel=RBM)
dbn.fit(Xtrain, pretrain_epochs=2)
if savefile:
dbn.save(savefile)
# first layer features
# initial weight is D x M
W = dbn.hidden_layers[0].W.eval()
for i in range(dbn.hidden_layers[0].M):
imgplot = plt.imshow(W[:, i].reshape(28, 28), cmap='gray')
plt.show()
should_quit = input("Show more? Enter 'n' to quit\n")
if should_quit == 'n':
break
# features learned in the last layer
for k in range(dbn.hidden_layers[-1].M):
# activate the kth node
X = dbn.fit_to_input(k)
imgplot = plt.imshow(X.reshape(28, 28), cmap='gray')
plt.show()
if k < dbn.hidden_layers[-1].M - 1:
should_quit = input("Show more? Enter 'n' to quit\n")
if should_quit == 'n':
break
def test_single_autoencoder():
Xtrain, Ytrain, Xtest, Ytest = getKaggleMNIST()
autoencoder = AutoEncoder(300, 0)
autoencoder.fit(Xtrain, epochs=2, show_fig=True)
done = False
while not done:
i = np.random.choice(len(Xtest))
x = Xtest[i]
y = autoencoder.predict([x])
plt.subplot(1, 2, 1)
plt.imshow(x.reshape(28, 28), cmap='gray')
plt.title('Original')
plt.subplot(1, 2, 2)
plt.imshow(y.reshape(28, 28), cmap='gray')
plt.title('Reconstructed')
plt.show()
ans = input("Generate another?")
if ans and ans[0] in ('n' or 'N'):
done = True
def main(loadfile=None, savefile=None):
Xtrain, Ytrain, Xtest, Ytest = getKaggleMNIST()
if loadfile:
dbn = DBN.load(loadfile)
else:
dbn = DBN([1000, 750, 500, 10]) # AutoEncoder is default
dbn = DBN([1000, 750, 500, 10], UnsupervisedModel=RBM)
dbn.fit(Xtrain, pretrain_epochs=15)
if savefile:
dbn.save(savefile)
# first layer features
# initial weight is D x M
# W = dbn.hidden_layers[0].W.eval()
# for i in xrange(dbn.hidden_layers[0].M):
# imgplot = plt.imshow(W[:,i].reshape(28, 28), cmap='gray')
# plt.show()
# should_quit = raw_input("Show more? Enter 'n' to quit\n")
# if should_quit == 'n':
# break
# features learned in the last layer
for k in xrange(dbn.hidden_layers[-1].M):
# activate the kth node
X = dbn.fit_to_input(k)
imgplot = plt.imshow(X.reshape(28, 28), cmap='gray')
plt.show()
if k < dbn.hidden_layers[-1].M - 1:
should_quit = raw_input("Show more? Enter 'n' to quit\n")
if should_quit == 'n':
break
def main():
Xtrain, Ytrain, Xtest, Ytest = getKaggleMNIST()
dbn = DBN([1000, 750, 500], UnsupervisedModel=AutoEncoder)
# dbn = DBN([1000, 750, 500, 10])
output = dbn.fit(Xtrain, pretrain_epochs=2)
print "output.shape", output.shape
# sample before using t-SNE because it requires lots of RAM
sample_size = 600
tsne = TSNE()
reduced = tsne.fit_transform(output[:sample_size])
plt.scatter(reduced[:,0], reduced[:,1], s=100, c=Ytrain[:sample_size], alpha=0.5)
plt.title("t-SNE visualization")
plt.show()
# t-SNE on raw data
reduced = tsne.fit_transform(Xtrain[:sample_size])
plt.scatter(reduced[:,0], reduced[:,1], s=100, c=Ytrain[:sample_size], alpha=0.5)
plt.title("t-SNE visualization")
plt.show()
pca = PCA()
reduced = pca.fit_transform(output)
plt.scatter(reduced[:,0], reduced[:,1], s=100, c=Ytrain, alpha=0.5)
plt.title("PCA visualization")
plt.show()
def main():
Xtrain, Ytrain, Xtest, Ytest = getKaggleMNIST()
dbn = DBN([1000, 750, 500], UnsupervisedModel=AutoEncoder)
# dbn = DBN([1000, 750, 500, 10])
output = dbn.fit(Xtrain, pretrain_epochs=2)
print "output.shape", output.shape
# sample before using t-SNE because it requires lots of RAM
sample_size = 600
tsne = TSNE()
reduced = tsne.fit_transform(output[:sample_size])
plt.scatter(reduced[:,0], reduced[:,1], s=100, c=Ytrain[:sample_size], alpha=0.5)
plt.title("t-SNE visualization")
plt.show()
# t-SNE on raw data
reduced = tsne.fit_transform(Xtrain[:sample_size])
plt.scatter(reduced[:,0], reduced[:,1], s=100, c=Ytrain[:sample_size], alpha=0.5)
plt.title("t-SNE visualization")
plt.show()
pca = PCA()
reduced = pca.fit_transform(output)
plt.scatter(reduced[:,0], reduced[:,1], s=100, c=Ytrain, alpha=0.5)
plt.title("PCA visualization")
plt.show()
def test_single_autoencoder():
Xtrain, Ytrain, Xtest, Ytest = getKaggleMNIST()
Xtrain = Xtrain.astype(np.float32)
Xtest = Xtest.astype(np.float32)
_, D = Xtrain.shape
autoencoder = AutoEncoder(D, 300, 0)
init_op = tf.global_variables_initializer()
with tf.Session() as session:
session.run(init_op)
autoencoder.set_session(session)
autoencoder.fit(Xtrain, show_fig=True)
done = False
while not done:
i = np.random.choice(len(Xtest))
x = Xtest[i]
y = autoencoder.predict([x])
plt.subplot(1, 2, 1)
plt.imshow(x.reshape(28, 28), cmap='gray')
plt.title('Original')
plt.subplot(1, 2, 2)
plt.imshow(y.reshape(28, 28), cmap='gray')
plt.title('Reconstructed')
plt.show()
ans = input("Generate another?")
if ans and ans[0] in ('n' or 'N'):
done = True
def singleAE():
X, Y, _, _ = getKaggleMNIST()
N, D = X.shape
#print("D ", D)
AE = Autoencoder(D, 20)
AE.fit(X, plot_fig=True)
done = False
while not done:
i = np.random.choice(len(X))
x = X[i]
# print(x)
im = AE.predict([x]).reshape(28, 28)
plt.subplot(1, 2, 1)
plt.imshow(x.reshape(28, 28), cmap='gray')
plt.title("Original")
plt.subplot(1, 2, 2)
plt.imshow(im, cmap='gray')
plt.title("Reconstruction")
plt.show()
ans = input("Generate another?")
if ans and ans[0] in ('n' or 'N'):
done = True
def test_single_autoencoder():
Xtrain, Ytrain, Xtest, Ytest = getKaggleMNIST()
autoencoder = AutoEncoder(300, 0)
autoencoder.fit(Xtrain, learning_rate=0.5, epochs=3, show_fig=True)
# 假如這裡epoch=0 就只會initialize weights 但不做training
done = False
while not done:
i = np.random.choice(len(Xtest))
x = Xtest[
i] # 這裡取出來的x.shape = (784,), 是一個vector，但 input of predict() need to be a matrix (1, 784)
y = autoencoder.predict(
[x]) # 這裡為什麼要寫 [x]? 這樣才可以把shape轉成 (1, 784) 作為input
plt.subplot(1, 2, 1)
plt.imshow(x.reshape(28, 28), cmap='gray')
plt.title('Original')
plt.subplot(1, 2, 2)
plt.imshow(y.reshape(28, 28), cmap='gray')
plt.title('Reconstructed')
plt.show()
ans = input("Generate another?")
if ans and ans[0] in ('n' or 'N'): # 第一次看到的寫法 in ('n' or 'N') ， 還不太懂
done = True
def main():
Xtrain, Ytrain, Xtest, Ytest = getKaggleMNIST()
dae = DeepAutoEncoder([500, 300, 2])
dae.fit(Xtrain)
mapping = dae.map2center(Xtrain)
plt.scatter(mapping[:,0], mapping[:,1], c=Ytrain, s=100, alpha=0.5)
plt.show()
def main():
Xtrain, Ytrain, _, _ = getKaggleMNIST()
bay = BayesClassifier()
bay.fit(Xtrain, Ytrain)
# = bay.sample()
for k in range(bay.K):
sample = bay.sample_given_y(k).reshape(28, 28)
mean = bay.gaussians[k]['m'].reshape(28, 28)
plt.subplot(1, 2, 1)
plt.imshow(sample, cmap='gray')
plt.title("sample")
plt.subplot(1, 2, 2)
plt.imshow(mean, cmap='gray')
plt.title("mean")
plt.show()
# generate rdm sample
sample = bay.sample().reshape(28, 28)
plt.imshow(sample, cmap='gray')
plt.title("rdm sample")
plt.show()
def main():
Xtrain, Ytrain, Xtest, Ytest = getKaggleMNIST()
# dnn = DNN([1000, 750, 500])
# dnn.fit(Xtrain, Ytrain, Xtest, Ytest, epochs=3)
# vs
dnn = DNN([1000, 750, 500])
dnn.fit(Xtrain, Ytrain, Xtest, Ytest, pretrain=False, epochs=10)
def main():
Xtrain, Ytrain, Xtest, Ytest = getKaggleMNIST()
dae = DeepAutoEncoder([500, 300, 2])
dae.fit(Xtrain)
mapping = dae.map2center(Xtrain)
plt.scatter(mapping[:,0], mapping[:,1], c=Ytrain, s=100, alpha=0.5)
plt.show()
def main():
xtr, ytr, xts, yts = getKaggleMNIST()
PCA_ = PCA()
reduced = PCA_.fit_transform(xtr)
print(xtr[:])
print(xtr.shape)
plt.figure()
plt.scatter(xtr[:,0], xtr[:,1], s=70, alpha=0.7)
plt.title("Image")
plt.figure()
plt.scatter(reduced[:,0],reduced[:,1],s=70,c=ytr,alpha=0.5)
plt.title("Image transformed")
plt.show()
plt.plot(PCA_.explained_variance_ratio_) # i.e. eigenvalues
plt.title("Variance/Eigenvalues")
plt.show()
cumulative_variance = []
last = 0
for v in PCA_.explained_variance_ratio_:
cumulative_variance.append(last+v)
last = cumulative_variance[-1]
plt.figure()
plt.plot(cumulative_variance)
plt.title("Cumulative variance")
plt.show()
def test_single_autoencoder():
Xtrain, Ytrain, Xtest, Ytest = getKaggleMNIST()
Xtrain = Xtrain.astype(np.float32)
Xtest = Xtest.astype(np.float32)
_, D = Xtrain.shape
autoencoder = AutoEncoder(D, 300, 0)
init_op = tf.global_variables_initializer()
with tf.Session() as session:
session.run(init_op)
autoencoder.set_session(session)
autoencoder.fit(Xtrain, show_fig=True)
done = False
while not done:
i = np.random.choice(len(Xtest))
x = Xtest[i]
y = autoencoder.predict([x])
plt.subplot(1,2,1)
plt.imshow(x.reshape(28,28), cmap='gray')
plt.title('Original')
plt.subplot(1,2,2)
plt.imshow(y.reshape(28,28), cmap='gray')
plt.title('Reconstructed')
plt.show()
ans = input("Generate another?")
if ans and ans[0] in ('n' or 'N'):
done = True
def main():
Xtrain, Ytrain, Xtest, Ytest = getKaggleMNIST()
dae = DeepAutoEncoder([500, 300, 2])
start_time = time.time() # 時間戳記
dae.fit(Xtrain, show_fig=True)
end_time = time.time() # 時間戳記
mapping = dae.map2center(Xtrain)
plt.scatter(mapping[:, 0], mapping[:, 1], c=Ytrain, s=100, alpha=0.5)
print(end_time - start_time)
def main():
train_X, train_Y, _, _, = getKaggleMNIST()
sample = 1000
train_X = train_X[:sample]
train_Y = train_Y[:sample]
tsne = TSNE()
Z = tsne.fit_transform(train_X)
plt.scatter(Z[:, 0], Z[:, 1], s=100, c=train_Y, alpha=0.5)
plt.show()
def main():
Xtrain, Ytrain, _, _ = getKaggleMNIST()
sample_size = 1000
X = Xtrain[:sample_size]
Y = Ytrain[:sample_size]
tsne = TSNE()
Z = tsne.fit_transform(X)
plt.scatter(Z[:,0], Z[:,1], s=100, c=Y, alpha=0.5)
plt.show()
def test_pretraining_dnn():
Xtrain, Ytrain, Xtest, Ytest = getKaggleMNIST()
_, D = Xtrain.shape
K = len(set(Ytrain))
dnn = DNN(D, [1000, 750, 500], K)
init_op = tf.global_variables_initializer()
with tf.Session() as session:
session.run(init_op)
dnn.set_session(session)
dnn.fit(Xtrain, Ytrain, Xtest, Ytest, pretrain=True, epochs=10)
def main():
Xtrain, Ytrain, _, _ = getKaggleMNIST()
sample_size = 1000
X = Xtrain[:sample_size]
Y = Ytrain[:sample_size]
tsne = TSNE()
Z = tsne.fit_transform(X)
plt.scatter(Z[:,0], Z[:,1], s=100, c=Y, alpha=0.5)
plt.show()
def main():
train_X, train_Y, test_X, test_Y = getKaggleMNIST()
train_X = train_X.astype(np.float32)
test_X = test_X.astype(np.float32)
N, D = train_X.shape
K = len(set(train_Y))
model = DNN(D, [1000, 750, 500], K)
init = tf.global_variables_initializer()
with tf.Session() as session:
session.run(init)
model.set_session(session)
model.fit(train_X, train_Y, test_X, test_Y, pretrain=True, epochs=10)
def main():
Xtrain, Ytrain, Xtest, Ytest = getKaggleMNIST()
dae = DeepAutoEncoder([500, 300, 2]) # 500 300 2 300 500
dae.fit(Xtrain)
mapping = dae.map2center(Xtrain)
plt.scatter(mapping[:,0], mapping[:,1], c=Ytrain, s=100, alpha=0.5)
plt.show()
# purity measure from unsupervised machine learning pt 1
_, Rfull = gmm(X, 10, max_iter=30)
print("full purity:", purity(Y, Rfull))
_, Rreduced = plot_k_means(Z, 10, max_iter=30)
print("reduced purity:", purity(Y, Rreduced))
def main():
xtr, ytr, xts, yts = getKaggleMNIST()
index = 0
print(xtr[index].shape)
print(xtr[index])
im = xtr[index].reshape((28,28))
#plt.figure()
#map(lambda x:plt.scatter(x,np.linspace(0,1)),xtr[index])
plt.figure()
plt.imshow(im)
plt.title("Label {}".format(ytr[index]))
plt.show()
def main():
train_X, train_Y, test_X, test_Y = getKaggleMNIST()
train_X = train_X.astype(np.float32)
N, D = train_X.shape
model = DeepAutoEncoder(D, [500, 300, 2])
init = tf.global_variables_initializer()
with tf.Session() as session:
session.run(init)
model.set_session(session)
model.fit(train_X)
mapping = model.transform(train_X)
plt.scatter(mapping[:, 0], mapping[:, 1], c=train_Y, s=100, alpha=0.5)
plt.show()
def main():
X_train, Y_train, X_test, Y_test = getKaggleMNIST()
X_train = X_train.astype(np.float32)
X_test = X_test.astype(np.float32)
_, D = X_train.shape
K = len(set(Y_train))
dnn = DNN(D, [1000, 750, 500], K, UnsupervisedModel=RBM)
init_op = tf.global_variables_initializer()
with tf.Session() as session:
session.run(init_op)
dnn.set_session(session)
dnn.fit(X_train, Y_train, X_test, Y_test, pretrain=True, epochs=10)
def main():
Xtrain, Ytrain, Xtest, Ytest = getKaggleMNIST()
# same as autoencoder_tf.py
Xtrain = Xtrain.astype(np.float32)
Xtest = Xtest.astype(np.float32)
_, D = Xtrain.shape
K = len(set(Ytrain))
dnn = DNN(D, [1000, 750, 500], K, UnsupervisedModel=RBM)
init_op = tf.compat.v1.global_variables_initializer()
with tf.compat.v1.Session() as session:
session.run(init_op)
dnn.set_session(session)
dnn.fit(Xtrain, Ytrain, Xtest, Ytest, pretrain=True, epochs=10)
def main():
Xtrain, Ytrain, Xtest, Ytest = getKaggleMNIST()
# same as autoencoder_tf.py
Xtrain = Xtrain.astype(np.float32)
Xtest = Xtest.astype(np.float32)
_, D = Xtrain.shape
K = len(set(Ytrain))
dnn = DNN(D, [1000, 750, 500], K, UnsupervisedModel=RBM)
init_op = tf.global_variables_initializer()
with tf.Session() as session:
session.run(init_op)
dnn.set_session(session)
dnn.fit(Xtrain, Ytrain, Xtest, Ytest, pretrain=True, epochs=10)
def singleVAE():
# get data
X, _, _, _ = getKaggleMNIST()
#X = (X > 0.5).astype(np.float32)
N, D = X.shape
print(D)
# used 768 in tutorial
vae = VAE(784, [200, 100])
vae.fit(X)
# plot reconstruction
done = False
while not done:
i = np.random.choice(len(X))
x = X[i]
im = vae.posterior_predictive_sample([x]).reshape(28, 28)
plt.subplot(1, 2, 1)
plt.imshow(x.reshape(28, 28), cmap='gray')
plt.title("Original")
plt.subplot(1, 2, 2)
plt.imshow(im, cmap='gray')
plt.title("Sampled")
plt.show()
ans = input("Generate another?")
if ans and ans[0] in ('n' or 'N'):
done = True
# plot output from random samples in latent space
done = False
while not done:
im, probs = vae.prior_predictive_sample_with_probs()
im = im.reshape(28, 28)
probs = probs.reshape(28, 28)
plt.subplot(1, 2, 1)
plt.imshow(im, cmap='gray')
plt.title("Prior predictive sample")
plt.subplot(1, 2, 2)
plt.imshow(probs, cmap='gray')
plt.title("Prior predictive probs")
plt.show()
ans = input("Generate another?")
if ans and ans[0] in ('n' or 'N'):
done = True
def test_pretraining_dnn():
Xtrain, Ytrain, Xtest, Ytest = getKaggleMNIST()
# dnn = DNN([1000, 750, 500])
# dnn.fit(Xtrain, Ytrain, Xtest, Ytest, epochs=3)
# vs
Xtrain = Xtrain.astype(np.float32)
Xtest = Xtest.astype(np.float32)
_, D = Xtrain.shape
K = len(set(Ytrain))
dnn = DNN(D, [1000, 750, 500], K)
init_op = tf.global_variables_initializer()
with tf.Session() as session:
session.run(init_op)
dnn.set_session(session)
dnn.fit(Xtrain, Ytrain, Xtest, Ytest, pretrain=True, epochs=10)
def test_pretraining_dnn():
Xtrain, Ytrain, Xtest, Ytest = getKaggleMNIST()
# dnn = DNN([1000, 750, 500])
# dnn.fit(Xtrain, Ytrain, Xtest, Ytest, epochs=3)
# vs
Xtrain = Xtrain.astype(np.float32)
Xtest = Xtest.astype(np.float32)
_, D = Xtrain.shape
K = len(set(Ytrain))
dnn = DNN(D, [1000, 750, 500], K)
init_op = tf.compat.v1.global_variables_initializer()
with tf.compat.v1.Session() as session:
session.run(init_op)
dnn.set_session(session)
dnn.fit(Xtrain, Ytrain, Xtest, Ytest, pretrain=True, epochs=10)
def main(): train_X, train_Y, test_X, test_Y = getKaggleMNIST() pca = PCA() train_Z = pca.fit_transform(train_X) plt.scatter(train_Z[:,0], train_Z[:,1], s=100, c=train_Y, alpha=0.5) plt.show() plt.plot(pca.explained_variance_ratio_) plt.show() cumulative_variance = [] last_variance = 0 for v in pca.explained_variance_ratio_: cumulative_variance.append(last_variance + v) last_variance = cumulative_variance[-1] plt.plot(cumulative_variance) plt.show()
def main():
Xtrain, Ytrain, _, _ = getKaggleMNIST()
sample_size = 1000
X = Xtrain[:sample_size]
Y = Ytrain[:sample_size]
tsne = TSNE()
Z = tsne.fit_transform(X)
plt.scatter(Z[:, 0], Z[:, 1], s=100, c=Y, alpha=0.5)
plt.show()
# purity measure from unsupervised machine learning pt 1
_, Rfull = gmm(X, 10, max_iter=30, smoothing=10e-1)
print "full purity:", purity(Y, Rfull)
_, Rreduced = gmm(Z, 10, max_iter=30, smoothing=10e-1)
print "reduced purity:", purity(Y, Rreduced)
def main():
Xtrain, Ytrain, Xtest, Ytest = getKaggleMNIST()
dae = DeepAutoEncoder([500, 300, 2])
dae.fit(Xtrain)
mapping = dae.map2center(Xtrain)
plt.scatter(mapping[:, 0], mapping[:, 1], c=Ytrain, s=100, alpha=0.5)
plt.show()
# purity measure from unsupervised machine learning pt 1
gmm = GaussianMixture(n_components=10)
gmm.fit(Xtrain)
responsibilities_full = gmm.predict_proba(Xtrain)
print "full purity:", purity(Ytrain, responsibilities_full)
gmm.fit(mapping)
responsibilities_reduced = gmm.predict_proba(mapping)
print "reduced purity:", purity(Ytrain, responsibilities_reduced)
def main():
X, Y, Xt, Yt = getKaggleMNIST()
sample_size = 500
print(X[0:sample_size].shape, Y[0:sample_size].shape)
#plt.scatter(X[0:10,0],X[0:10,1], s=70,c=Y[0:10],alpha=.5)
#plt.show()
tsne = TSNE()
Z = tsne.fit_transform(X[0:sample_size], Y[:sample_size])
plt.subplot(211)
plt.scatter(Z[:, 0], Z[:, 1], s=70, c=Y[:sample_size], alpha=.5)
plt.colorbar()
Z1 = tsne.fit_transform(Z)
plt.subplot(212)
plt.scatter(Z1[:, 0], Z1[:, 1], s=70, c=Y[:sample_size], alpha=.5)
plt.colorbar()
plt.show()
def main():
Xtrain, Ytrain, _, _ = getKaggleMNIST()
sample_size = 1000
X = Xtrain[:sample_size]
Y = Ytrain[:sample_size]
tsne = TSNE()
Z = tsne.fit_transform(X)
plt.scatter(Z[:,0], Z[:,1], s=100, c=Y, alpha=0.5)
plt.show()
# purity measure from unsupervised machine learning pt 1
_, Rfull = gmm(X, 10, max_iter=30, smoothing=10e-1)
print "full purity:", purity(Y, Rfull)
_, Rreduced = gmm(Z, 10, max_iter=30, smoothing=10e-1)
print "reduced purity:", purity(Y, Rreduced)
def main():
Xtrain, Ytrain, Xtest, Ytest = getKaggleMNIST()
dae = DeepAutoEncoder([500, 300, 2])
dae.fit(Xtrain)
mapping = dae.map2center(Xtrain)
plt.scatter(mapping[:,0], mapping[:,1], c=Ytrain, s=100, alpha=0.5)
plt.show()
# purity measure from unsupervised machine learning pt 1
gmm = GaussianMixture(n_components=10)
gmm.fit(Xtrain)
responsibilities_full = gmm.predict_proba(Xtrain)
print "full purity:", purity(Ytrain, responsibilities_full)
gmm.fit(mapping)
responsibilities_reduced = gmm.predict_proba(mapping)
print "reduced purity:", purity(Ytrain, responsibilities_reduced)
def main():
X, _, _, _ = getKaggleMNIST()
X = (X > 0.5).astype(np.float32) # we do bernuli
vae = VariationalAutoencoder(784, [200, 100])
vae.fit(X)
# plot reconstruction
done = False
while not done:
i = np.random.chocie(len(X))
x = X[i]
im = vae.posterior_predictive_sample([x]).reshape(28, 28)
plt.subplot(1, 2, 1)
plt.imshow(x.reshape(28, 28), cmap='gray')
plt.title("Originial")
plt.subplot(1, 2, 2)
plt.imshow(im, cmap='gray')
plt.title("Sampled")
plt.show()
# ans = input("Gen next one?")
# if ans and ans[0] in ('n' or 'N'):
# done = True
done = True
# plot reconstruction
done = False
while not done:
im, probs = vae.prior_predictive_sample_with_probs()
im = im.reshape(28, 28)
probs = probs.reshape(28, 28)
plt.subplot(1, 2, 1)
plt.imshow(im, cmap='gray')
plt.title("Originial")
plt.subplot(1, 2, 2)
plt.imshow(probs, cmap='gray')
plt.title("Sampled")
plt.show()
# ans = input("Gen next one?")
# if ans and ans[0] in ('n' or 'N'):
# done = True
done = True
def main(): Xtrain, Ytrain, Xtest, Ytest = getKaggleMNIST() pca = PCA() reduced = pca.fit_transform(Xtrain) plt.scatter(reduced[:,0], reduced[:,1], s=100, c=Ytrain, alpha=0.5) plt.show() plt.plot(pca.explained_variance_ratio_) plt.show() cumulative = [] last = 0 for v in pca.explained_variance_ratio_: cumulative.append(last + v) last = cumulative[-1] plt.plot(cumulative) plt.show()
def main():
Xtrain, Ytrain, Xtest, Ytest = getKaggleMNIST()
dae = DeepAutoEncoder([500, 300, 2])
dae.fit(Xtrain)
mapping = dae.map2center(Xtrain)
plt.scatter(mapping[:,0], mapping[:,1], c=Ytrain, s=100, alpha=0.5)
plt.show()
# purity measure from unsupervised machine learning pt 1
# NOTE: this will take a long time (i.e. just leave it overnight)
gmm = GaussianMixture(n_components=10)
gmm.fit(Xtrain)
print("Finished GMM training")
responsibilities_full = gmm.predict_proba(Xtrain)
print("full purity:", purity(Ytrain, responsibilities_full))
gmm.fit(mapping)
responsibilities_reduced = gmm.predict_proba(mapping)
print("reduced purity:", purity(Ytrain, responsibilities_reduced))
def main():
Xtrain, Ytrain, Xtest, Ytest = getKaggleMNIST()
dae = DeepAutoEncoder([500, 300, 2])
dae.fit(Xtrain)
mapping = dae.map2center(Xtrain)
plt.scatter(mapping[:, 0], mapping[:, 1], c=Ytrain, s=100, alpha=0.5)
plt.show()
# purity measure from unsupervised machine learning pt 1
# NOTE: this will take a long time (i.e. just leave it overnight)
gmm = GaussianMixture(n_components=10)
gmm.fit(Xtrain)
print("Finished GMM training")
responsibilities_full = gmm.predict_proba(Xtrain)
print("full purity:", purity(Ytrain, responsibilities_full))
gmm.fit(mapping)
responsibilities_reduced = gmm.predict_proba(mapping)
print("reduced purity:", purity(Ytrain, responsibilities_reduced))
def main():
Xtrain, Ytrain, Xtest, Ytest = getKaggleMNIST()
pca = PCA()
reduced = pca.fit_transform(Xtrain)
plt.scatter(reduced[:,0], reduced[:,1], s=100, c=Ytrain, alpha=0.5)
plt.show()
plt.plot(pca.explained_variance_ratio_)
plt.show()
# cumulative variance
# choose k = number of dimensions that gives us 95-99% variance
cumulative = []
last = 0
for v in pca.explained_variance_ratio_:
cumulative.append(last + v)
last = cumulative[-1]
plt.plot(cumulative)
plt.show()
def main():
Xtrain, Ytrain, Xtest, Ytest = getKaggleMNIST()
pca = PCA()
reduced = pca.fit_transform(Xtrain)
plt.scatter(reduced[:, 0], reduced[:, 1], s=100, c=Ytrain, alpha=0.5)
plt.show()
plt.plot(pca.explained_variance_ratio_)
plt.show()
# cumulative variance
# choose k = number of dimensions that gives us 95-99% variance
cumulative = []
last = 0
for v in pca.explained_variance_ratio_:
cumulative.append(last + v)
last = cumulative[-1]
plt.plot(cumulative)
plt.show()
def test_single_autoencoder():
Xtrain, Ytrain, Xtest, Ytest = getKaggleMNIST()
autoencoder = AutoEncoder(300, 0)
autoencoder.fit(Xtrain, epochs=2, show_fig=True)
done = False
while not done:
i = np.random.choice(len(Xtest))
x = Xtest[i]
y = autoencoder.predict([x])
plt.subplot(1,2,1)
plt.imshow(x.reshape(28,28), cmap='gray')
plt.title('Original')
plt.subplot(1,2,2)
plt.imshow(y.reshape(28,28), cmap='gray')
plt.title('Reconstructed')
plt.show()
ans = input("Generate another?")
if ans and ans[0] in ('n' or 'N'):
done = True
def main():
Xtrain, Ytrain, Xtest, Ytest = getKaggleMNIST()
dnn = DNN([1000, 750, 500], UnsupervisedModel=RBM)
dnn.fit(Xtrain, Ytrain, Xtest, Ytest, epochs=3)
def main():
Xtrain, Ytrain, Xtest, Ytest = getKaggleMNIST()
dnn = ANN([1000, 750, 500])
dnn.fit(Xtrain, Ytrain)
# https://www.udemy.com/data-science-deep-learning-in-theano-tensorflow
from __future__ import print_function, division
from future.utils import iteritems
from builtins import range, input
# Note: you may need to update your version of future
# sudo pip install -U future
import numpy as np
from sklearn.neural_network import MLPRegressor
from util import getKaggleMNIST
# get data
X, _, Xt, _ = getKaggleMNIST()
# create the model and train it
model = MLPRegressor()
model.fit(X, X)
# test the model
print("Train R^2:", model.score(X, X))
print("Test R^2:", model.score(Xt, Xt))
Xhat = model.predict(X)
mse = ((Xhat - X)**2).mean()
print("Train MSE:", mse)
Xhat = model.predict(Xt)
mse = ((Xhat - Xt)**2).mean()
# https://deeplearningcourses.com/c/unsupervised-deep-learning-in-python # https://www.udemy.com/unsupervised-deep-learning-in-python from __future__ import print_function, division from builtins import range, input # Note: you may need to update your version of future # sudo pip install -U future import numpy as np import matplotlib.pyplot as plt from sklearn.decomposition import PCA, TruncatedSVD from util import getKaggleMNIST X, Y, _, _ = getKaggleMNIST() m = X.mean(axis=0) s = X.std(axis=0) np.place(s, s == 0, 1) X = (X - m) / s pca = PCA() svd = TruncatedSVD() Z1 = pca.fit_transform(X) Z2 = svd.fit_transform(X) plt.subplot(1,2,1) plt.scatter(Z1[:,0], Z1[:,1], c=Y) plt.subplot(1,2,2) plt.scatter(Z2[:,0], Z2[:,1], c=Y) plt.show()
from __future__ import print_function, division from builtins import range # Note: you may need to update your version of future # sudo pip install -U future import numpy as np import matplotlib.pyplot as plt from util import getKaggleMNIST from keras.models import Model from keras.layers import Dense, Activation, Input, Dropout # get the data Xtrain, Ytrain, Xtest, Ytest = getKaggleMNIST() # get shapes N, D = Xtrain.shape K = len(set(Ytrain)) # ANN with layers [784] -> [500] -> [300] -> [10] i = Input(shape=(D,)) x = Dropout(0.2)(i) x = Dense(500, activation='relu')(x) x = Dropout(0.5)(x) x = Dense(300, activation='relu')(x) x = Dropout(0.5)(x) x = Dense(K, activation='softmax')(x) # instantiate the model object
