def train_initial(name, n_seq, n_labels, n_dimension, n_hidden_1, n_hidden_2, epochs, save):
usage_ratio = 1
# epochs = 150
print '========================= Reading ========================='
X_train, y_train, X_test, y_test = test.read_data(name=name, n_seq=n_seq, n_labels=n_labels, n_dimension=n_dimension)
data = (X_train, y_train, X_test, y_test)
print '========================= Modeling ========================='
model = lstm.build_model(n_dimension=n_dimension, n_labels=n_labels, n_seq=n_seq, n_hidden_1=n_hidden_1, n_hidden_2=n_hidden_2)
print '========================= Training =========================='
model = lstm.run_network(model=model, data=data, epochs=epochs, usage_ratio=usage_ratio, save=True, save_name=name)
print '========================= Testing =========================='
test.test_all_metrics(model, data=data, usage_ratio=usage_ratio)
def get_image_for_frame(i):
# r=np.zeros((height,width))
# r[:,i]=1.0
# r[:, -i] = 0.5
# return r
files = test.list_files() #list all .aps files in sequential order
r = files[
0] #0 stands for subject 0, all 16 images for this subject are included
header = test.read_header(r) #extracting the header dictionary
r = test.read_data(
r) #reading the binary .aps file & extracting image data
r = r.transpose() #so we can index each of the 16 images
r = r[i] #images 1-16
print(r.max())
print(r.min())
print(r.mean())
r = r.astype(np.double)
print(r.shape)
r = r[::-1, ]
#r=(test.get_single_image(files[0],0))[i].flatten()
#r=r.astype(np.double)
return r / r.max()
from sklearn import svm, datasets
import numpy as np
from test import read_data
import pylab as pl
filename = 'svm_train_2.svm'
points, labels = read_data(filename)
for i in range(labels.shape[0]):
if labels[i] == -1:
labels[i] = 0
iris = datasets.load_iris()
X = iris.data[:,:2]
y = iris.target
clf = svm.SVC(kernel = 'linear')
clf.fit(X,y)
# create mesh
h = 0.02
xMin, xMax = X[:,0].min()-1, X[:,0].max()+1
yMin, yMax = X[:,1].min()-1, X[:,1].max()+1
xx, yy = np.meshgrid(np.arange(xMin, xMax, h),
np.arange(yMin, yMax, h))
Z = clf.predict(np.c_[xx.ravel(),yy.ravel()])
Z = Z.reshape(xx.shape)
pl.contourf(xx,yy,Z,cmap=pl.cm.Paired)
pl.axis('off')
pl.scatter(X[:,0],X[:,1], c = y, cmap=pl.cm.Paired)
pl.title(filename)
pl.show()
from sklearn import svm, datasets
import numpy as np
from test import read_data
import pylab as pl
filename = 'svm_train_2.svm'
points, labels = read_data(filename)
for i in range(labels.shape[0]):
if labels[i] == -1:
labels[i] = 0
iris = datasets.load_iris()
X = iris.data[:, :2]
y = iris.target
clf = svm.SVC(kernel='linear')
clf.fit(X, y)
# create mesh
h = 0.02
xMin, xMax = X[:, 0].min() - 1, X[:, 0].max() + 1
yMin, yMax = X[:, 1].min() - 1, X[:, 1].max() + 1
xx, yy = np.meshgrid(np.arange(xMin, xMax, h), np.arange(yMin, yMax, h))
Z = clf.predict(np.c_[xx.ravel(), yy.ravel()])
Z = Z.reshape(xx.shape)
pl.contourf(xx, yy, Z, cmap=pl.cm.Paired)
pl.axis('off')
pl.scatter(X[:, 0], X[:, 1], c=y, cmap=pl.cm.Paired)
pl.title(filename)
pl.show()
def run_network(save_name, model=None, data=None, epochs=100, usage_ratio=1, save=False):
epochs = epochs
batch_size = 128
start_time = time.time()
print 'Running LSTM, start time: ', start_time
if data is None:
# X_train, y_train, X_test, y_test = divide_data(usage_ratio=usage_ratio)
X_train, y_train, X_test, y_test = read_data(train_name='train_2.csv', test_name='test_2.csv', usage_ratio=usage_ratio)
else:
X_train, y_train, X_test, y_test = data
print 'Modeling with ', X_train.shape[0], ' training samples & ', \
X_test.shape[0], ' testing samples..'
if model is None:
model = build_model()
# earlyStopping=EarlyStopping(monitor='val_loss',
# patience=5,
# verbose=1,
# mode='auto')
print "Training Mode: TENSORBOARD ON.."
try:
model.fit(
X_train, y_train,
batch_size=batch_size,
nb_epoch=epochs,
validation_data=(X_test, y_test),
# using TB to visualize process.
callbacks=[TensorBoard(log_dir='../logs')]
)
except KeyboardInterrupt:
print 'Training over, total executing time: ', time.time() - start_time
print 'Training over, total executing time: ', time.time() - start_time
# y_train_pred = model.predict(X_train)
# y_test_pred = model.predict(X_test)
if save:
if raw_input('Save the model? (y/n): ') == 'y':
json_string = model.to_json()
filename1 = 'lstm_architecture'
filename2 = 'lstm_weights'
cnt = 1
model_path = '../models/'
for root, dirs, files in os.walk(model_path):
for filename in files:
if 'lstm_architecture' in filename:
cnt += 1
filename1 = model_path + filename1 + '_' + save_name + '.json'
filename2 = model_path + filename2 + '_' + save_name + '.h5'
print 'Saving LSTM model into .json & .h5 at:', filename1
open(filename1, 'w').write(json_string)
model.save_weights(filename2)
acc_train = model.evaluate(X_train, y_train, batch_size=128)
acc_test = model.evaluate(X_test, y_test, batch_size=128)
print '\ntraining accuracy: ', acc_train
print 'testing accuracy: ', acc_test
print '\nTraining down.'
return model