def chain_crf():
letters = load_letters()
x, y, folds = letters['data'], letters['labels'], letters['folds']
print "Letters : "
print letters
# print "Data : "
# print letters['data']
# print "Labels : "
# print letters['labels']
x, y = np.array(x), np.array(y)
x_train, x_test = x[folds == 1], x[folds != 1]
y_train, y_test = y[folds == 1], y[folds != 1]
print len(x_train)
print len(x_test)
print "Done"
print x_train[0].shape
print y_train[0].shape
print x_train[10].shape
print y_train[10].shape
model = ChainCRF()
ssvm = FrankWolfeSSVM(model=model, C=.1, max_iter=10)
print ssvm.fit(x_train, y_train)
print ssvm.score(x_test, y_test)
def load_data(self):
alphabets = load_letters()
X = np.array(alphabets['data'])
y = np.array(alphabets['labels'])
folds = alphabets['folds']
return X, y, folds
def __init__(self):
letters = load_letters()
X, y, folds = letters['data'], letters['labels'], letters['folds']
self.dict = "abcdefghijklmnopqrstuvwxyz"
X, y = np.array(X), np.array(y)
self.X_train, self.X_test = X[folds == 1], X[folds != 1]
self.y_train, self.y_test = y[folds == 1], y[folds != 1]
# shuffle
num_ex = len(self.X_train)
perm = np.arange(num_ex)
np.random.shuffle(perm)
self.X_train = self.X_train[perm]
self.y_train = self.y_train[perm]
self.DIM = self.X_train[0].shape[1]
self.test_num = len(self.X_test)
self.train_num = len(self.X_train)
self.y_one_hot_train = self.to_one_hot(self.y_train)
self.y_one_hot_test = self.to_one_hot(self.y_test)
self.X_lst_train, self.y_lst_train = self._get_lst(self.X_train, self.y_train)
self.X_lst_test, self.y_lst_test = self._get_lst(self.X_test, self.y_test)
def chain_crf():
from pystruct.datasets import load_letters
letters = load_letters()
X, y, folds = ut.take(letters, ['data', 'labels', 'folds'])
X, y = np.array(X), np.array(y)
X_train, X_test = X[folds == 1], X[folds != 1]
y_train, y_test = y[folds == 1], y[folds != 1]
len(X_train)
def chain_crf():
from pystruct.datasets import load_letters
letters = load_letters()
X, y, folds = ut.take(letters, ['data', 'labels', 'folds'])
X, y = np.array(X), np.array(y)
X_train, X_test = X[folds == 1], X[folds != 1]
y_train, y_test = y[folds == 1], y[folds != 1]
len(X_train)
def loadWindows(start, stop, window):
letters = load_letters()
X, y = letters['data'], letters['labels']
X, y = np.array(X), np.array(y)
word_features = selectWordsByLetters(X, start, stop)
word_labels = selectWordsByLetters(y, start, stop)
window_features, window_labels = splitWords(word_features, word_labels,
window)
return (window_features.astype(np.double), window_labels.astype(np.double))
There are obvious extensions that both methods could benefit from, such as window features or non-linear kernels. This example is more meant to give a demonstration of the CRF than to show its superiority. """ import numpy as np import matplotlib.pyplot as plt from sklearn.svm import LinearSVC from pystruct.datasets import load_letters from pystruct.models import ChainCRF from pystruct.learners import FrankWolfeSSVM abc = "abcdefghijklmnopqrstuvwxyz" letters = load_letters() X, y, folds = letters['data'], letters['labels'], letters['folds'] # we convert the lists to object arrays, as that makes slicing much more # convenient X, y = np.array(X), np.array(y) X_train, X_test = X[folds == 1], X[folds != 1] y_train, y_test = y[folds == 1], y[folds != 1] # Train linear SVM svm = LinearSVC(dual=False, C=.1) # flatten input svm.fit(np.vstack(X_train), np.hstack(y_train)) # Train linear chain CRF model = ChainCRF() ssvm = FrankWolfeSSVM(model=model, C=.1, max_iter=11)
import numpy as np
import matplotlib.pyplot as plt
import os
from sklearn.svm import LinearSVC
from sklearn.svm import SVC
from common.viewers.imshow import imshow
from pystruct.datasets import load_letters
from pystruct.models import ChainCRF, GraphCRF
from pystruct.learners import FrankWolfeSSVM
from sklearn.linear_model import LinearRegression
from common.utils import get_letters_in_pred_like, arrange_letters_in_pred_like
import cPickle
abc = "abcdefghijklmnopqrstuvwxyz"
# Load data:
letters = load_letters()
X, y, folds = letters['data'], letters['labels'], letters['folds']
# we convert the lists to object arrays, as that makes slicing much more
# convenient
X, y = np.array(X), np.array(y)
X_train, X_test = X[folds == 1], X[folds != 1]
y_train, y_test = y[folds == 1], y[folds != 1]
net_base_path = '/media/ohadsh/sheard/googleDrive/Master/courses/probabilistic_graphical_models/outputs/part_3/training_2016_06_11/'
# Load pre-trained network
train_name = 'train_pred_-1.pkl'
test_name = 'test_pred_-1.pkl'
with open(os.path.join(net_base_path, train_name), 'r') as f:
train_net_pred = cPickle.load(f)
def load_data(self):
letters = load_letters()
X, y, folds = letters['data'], letters['labels'], letters['folds']
X, y = np.array(X), np.array(y)
return X, y, folds
def load_data(self):
letters = load_letters()
X, y, folds = letters['data'], letters['labels'], letters['folds']
X, y = np.array(X), np.array(y)
return X, y, folds
def test_dataset_loading():
# test that we can read the datasets.
load_scene()
load_letters()
load_snakes()
