def get_model(self):
# Don't need to build the model again, after loading it.
if self.load_model:
self.model = load_object_from_file(self.load_model)
self.skip_examples = self.model.total_examples
# Don't use the stored learning rate, but the one the user provided.
# Problem with that: Sometimes we don't want that. Especially when
# the model should proceed the training with learning rate it has
# when it stopped training.
# self.model.set_learning_rate(float(self.learning_rate),
# self.learning_method, self.lr_adaptation_method)
if self.period_type == 'time':
self._last_validation = time()
else:
self._last_validation = self.model.total_examples
else:
self.model = self.create_model()
self.model.set_learning_rate(self.learning_rate,
self.learning_method,
self.lr_adaptation_method)
if self.load_params is not None:
self.model.load_params(self.load_params[0], self.load_params[1])
self.model.link(self.inputs)
self.model.build()
def get_model(self):
# Don't need to build the model again, after loading it.
if self.load_model:
self.model = load_object_from_file(self.load_model)
self.skip_examples = self.model.total_examples
# Don't use the stored learning rate, but the one the user provided.
# Problem with that: Sometimes we don't want that. Especially when
# the model should proceed the training with learning rate it has
# when it stopped training.
# self.model.set_learning_rate(float(self.learning_rate),
# self.learning_method, self.lr_adaptation_method)
if self.period_type == 'time':
self._last_validation = time()
else:
self._last_validation = self.model.total_examples
else:
self.model = self.create_model()
self.model.set_learning_rate(self.learning_rate,
self.learning_method, self.lr_adaptation_method)
if self.load_params is not None:
self.model.load_params(self.load_params[0], self.load_params[1])
self.model.link(self.inputs)
self.model.build()
def main(argv=None):
if argv is None:
argv = sys.argv[1:]
args = parser.parse_args(argv)
log.info('start parameters: ' + str(args))
log.info('loading data')
model = load_object_from_file(args.model_file)
log.info('writing data')
# trainer.dump_vocabulary(args.vocabulary_file)
model.store_params(args.store_params[0], args.store_params[1], True,
args.format)
log.info('finished')
def main(argv=None):
if argv is None:
argv = sys.argv[1:]
args = parser.parse_args(argv)
log.info('start parameters: ' + str(args))
log.info('loading data')
model = load_object_from_file(args.model_file)
log.info('writing data')
# trainer.dump_vocabulary(args.vocabulary_file)
model.store_params(args.store_params[0], args.store_params[1], True,
args.format)
log.info('finished')
def main(argv=None):
if argv is None:
argv = sys.argv[1:]
args = parser.parse_args(argv)
log.info('start parameters: ' + str(args))
log.info('loading data')
model = load_object_from_file(args.model_file)
# read vocabulary from file
vocab = sort_dict_by_label(read_vocabulary_id_file(args.vocabulary))
# get matrices from model
r_matrix = model.R.get_value()
q_matrix = model.Q.get_value()
# get input embeddings
if args.model_type == 'vlbl':
in_we = r_matrix
elif args.model_type == 'vlbl_dist':
# this will not work with the old versions of models - because of sparsity
d_matrix = model.D.get_value().todense()
in_we = np.dot(d_matrix, r_matrix)
# need to convert from numpy.matrix to numpy.ndarray
in_we = in_we.view(type=np.ndarray)
with utf8_file_open(args.result_file + ".in", 'w') as outfile:
for (word, ind) in vocab:
outfile.write(
unicode(word) + u' ' + u' '.join(map(str, in_we[ind])) + u'\n')
with utf8_file_open(args.result_file + ".out", 'w') as outfile:
for (word, ind) in vocab:
outfile.write(
unicode(word) + u' ' + u' '.join(map(str, q_matrix[ind])) +
u'\n')
log.info('finished')
def main(argv=None):
if argv is None:
argv = sys.argv[1:]
args = parser.parse_args(argv)
log.info('start parameters: ' + str(args))
log.info('loading data')
model = load_object_from_file(args.model_file)
# read vocabulary from file
vocab = sort_dict_by_label(read_vocabulary_id_file(args.vocabulary))
# get matrices from model
r_matrix = model.R.get_value()
q_matrix = model.Q.get_value()
# get input embeddings
if args.model_type == 'vlbl':
in_we = r_matrix
elif args.model_type == 'vlbl_dist':
# this will not work with the old versions of models - because of sparsity
d_matrix = model.D.get_value().todense()
in_we = np.dot(d_matrix, r_matrix)
# need to convert from numpy.matrix to numpy.ndarray
in_we = in_we.view(type=np.ndarray)
with utf8_file_open(args.result_file + ".in", 'w') as outfile:
for (word, ind) in vocab:
outfile.write(unicode(word) + u' ' + u' '.join(map(str, in_we[ind])) + u'\n')
with utf8_file_open(args.result_file + ".out", 'w') as outfile:
for (word, ind) in vocab:
outfile.write(unicode(word) + u' ' + u' '.join(map(str, q_matrix[ind])) + u'\n')
log.info('finished')
def get_model(self):
self.model = load_object_from_file(self.load_model)
self.predictor_method = self.model.predictor
for i in range(self.training_steps):
print((self.b.get_value(), type(self.b.get_value())))
pred, err = self.train(self.D[0], self.D[1])
print(floatX)
if not os.path.exists('test'):
print('create new model')
model = LogisticRegression()
model.create_graph()
model.do_train()
save_object_to_file(model, 'test')
else:
print('load model')
model = load_object_from_file('test')
print("Final model:")
print((model.w.get_value(), model.b.get_value()))
print(("target values for D:", model.D[1]))
print(("prediction on D:", model.predict(model.D[0])))
# import numpy
# import theano
# import theano.tensor as T
# from theano import config
# rng = numpy.random
# floatX = config.floatX
#
# N = 400
def get_model(self):
self.model = load_object_from_file(self.load_model)
self.predictor_method = self.model.predictor
for i in range(self.training_steps):
print self.b.get_value(), type(self.b.get_value())
pred, err = self.train(self.D[0], self.D[1])
print floatX
if not os.path.exists('test'):
print 'create new model'
model = LogisticRegression()
model.create_graph()
model.do_train()
save_object_to_file(model, 'test')
else:
print 'load model'
model = load_object_from_file('test')
print "Final model:"
print model.w.get_value(), model.b.get_value()
print "target values for D:", model.D[1]
print "prediction on D:", model.predict(model.D[0])
# import numpy
# import theano
# import theano.tensor as T
# from theano import config
# rng = numpy.random
# floatX = config.floatX
#
# N = 400