def test_save_load(self):
# Data
x_batch = np.random.randint(0, VOCABULARY_SIZE, [BATCH_SIZE, SEQUENCE_LENGTH])
y_batch = np.eye(2)[np.ones(BATCH_SIZE, dtype=np.int32)]
_, checkpoint_path = tempfile.mkstemp()
with self.graph.as_default(), self.sess.as_default():
cnn = self._build_classifier()
self.sess.run(tf.initialize_all_variables())
saver = tf.train.Saver()
feed_dict = {cnn.input_x: x_batch, cnn.input_y: y_batch}
loss1 = self.sess.run(cnn.total_loss, feed_dict)
saver.save(self.sess, checkpoint_path)
# Save model hyperparameters
f, path = tempfile.mkstemp()
cnn.save_to_file(path)
with tf.Graph().as_default(), tf.Session() as sess:
cnn2 = CNNClassifier.load_from_file(path)
x = tf.placeholder(tf.int32, [BATCH_SIZE, SEQUENCE_LENGTH])
y = tf.placeholder(tf.float32, [BATCH_SIZE, 2])
cnn2.build_graph(x, y)
saver = tf.train.Saver()
saver.restore(sess, checkpoint_path)
feed_dict = {cnn2.input_x: x_batch, cnn2.input_y: y_batch}
loss2 = sess.run(cnn2.total_loss, feed_dict)
self.assertEqual(loss1, loss2)
def _build_classifier(self):
with self.graph.as_default(), self.sess.as_default():
tf.set_random_seed(42)
x = tf.placeholder(tf.int32, [None, SEQUENCE_LENGTH])
y = tf.placeholder(tf.float32, [None, 2])
cnn = CNNClassifier(sequence_length=SEQUENCE_LENGTH,
vocabulary_size=VOCABULARY_SIZE,
num_classes=2,
embedding_dim=64,
filter_sizes="2,3,4",
num_filters=10,
use_highway=False,
dropout_keep_prob_embedding=1.0,
dropout_keep_prob_features=1.0)
cnn.build_graph(x, y)
return cnn
def test_save_load(self):
# Data
x_batch = np.random.randint(0, VOCABULARY_SIZE,
[BATCH_SIZE, SEQUENCE_LENGTH])
y_batch = np.eye(2)[np.ones(BATCH_SIZE, dtype=np.int32)]
_, checkpoint_path = tempfile.mkstemp()
with self.graph.as_default(), self.sess.as_default():
cnn = self._build_classifier()
self.sess.run(tf.initialize_all_variables())
saver = tf.train.Saver()
feed_dict = {cnn.input_x: x_batch, cnn.input_y: y_batch}
loss1 = self.sess.run(cnn.total_loss, feed_dict)
saver.save(self.sess, checkpoint_path)
# Save model hyperparameters
f, path = tempfile.mkstemp()
cnn.save_to_file(path)
with tf.Graph().as_default(), tf.Session() as sess:
cnn2 = CNNClassifier.load_from_file(path)
x = tf.placeholder(tf.int32, [BATCH_SIZE, SEQUENCE_LENGTH])
y = tf.placeholder(tf.float32, [BATCH_SIZE, 2])
cnn2.build_graph(x, y)
saver = tf.train.Saver()
saver.restore(sess, checkpoint_path)
feed_dict = {cnn2.input_x: x_batch, cnn2.input_y: y_batch}
loss2 = sess.run(cnn2.total_loss, feed_dict)
self.assertEqual(loss1, loss2)
def _build_classifier(self):
with self.graph.as_default(), self.sess.as_default():
tf.set_random_seed(42)
x = tf.placeholder(tf.int32, [None, SEQUENCE_LENGTH])
y = tf.placeholder(tf.float32, [None, 2])
cnn = CNNClassifier(
sequence_length=SEQUENCE_LENGTH,
vocabulary_size=VOCABULARY_SIZE,
num_classes=2,
embedding_dim=64,
filter_sizes="2,3,4",
num_filters=10,
use_highway=False,
dropout_keep_prob_embedding=1.0,
dropout_keep_prob_features=1.0)
cnn.build_graph(x, y)
return cnn
from tfmodels.data.mr import MRData
from tfmodels.data.imdb import IMDBData
from tfmodels.data.ymrjp import YMRJPData
from datetime import datetime
from tfmodels.models.cnn.cnn_classifier import CNNClassifier, CNNClassifierTrainer, CNNClassifierEvaluator
pp = pprint.PrettyPrinter(indent=2)
# Pick training data
tf.flags.DEFINE_boolean("data_mr", False, "Dataset: MR")
tf.flags.DEFINE_boolean("data_sst", False, "Dataset: SST")
tf.flags.DEFINE_boolean("data_ymrjp", False,
"Dataset: Yahoo Movie Reviews (Japanese)")
# Classifier parameters
CNNClassifier.add_flags()
# Training parameters
tf.flags.DEFINE_integer("batch_size", 64,
"Size for one batch of training/dev examples")
tf.flags.DEFINE_integer("random_state", 42,
"Random state initialization for reproducibility")
tf.flags.DEFINE_integer("max_sequence_length", 512,
"Examples will be padded/truncated to this length")
tf.flags.DEFINE_integer("num_epochs", 20, "Number of training epochs")
tf.flags.DEFINE_integer(
"evaluate_every", 50,
"Evaluate model on dev set after this number of steps")
tf.flags.DEFINE_integer(
"checkpoint_every", 50,
"Evaluate model on dev set after this number of steps")
import tfmodels.data.utils
from tfmodels.data.mr import MRData
from tfmodels.data.imdb import IMDBData
from tfmodels.data.ymrjp import YMRJPData
from datetime import datetime
from tfmodels.models.cnn.cnn_classifier import CNNClassifier, CNNClassifierTrainer, CNNClassifierEvaluator
pp = pprint.PrettyPrinter(indent=2)
# Pick training data
tf.flags.DEFINE_boolean("data_mr", False, "Dataset: MR")
tf.flags.DEFINE_boolean("data_sst", False, "Dataset: SST")
tf.flags.DEFINE_boolean("data_ymrjp", False, "Dataset: Yahoo Movie Reviews (Japanese)")
# Classifier parameters
CNNClassifier.add_flags()
# Training parameters
tf.flags.DEFINE_integer("batch_size", 64, "Size for one batch of training/dev examples")
tf.flags.DEFINE_integer("random_state", 42, "Random state initialization for reproducibility")
tf.flags.DEFINE_integer("max_sequence_length", 512, "Examples will be padded/truncated to this length")
tf.flags.DEFINE_integer("num_epochs", 20, "Number of training epochs")
tf.flags.DEFINE_integer("evaluate_every", 50, "Evaluate model on dev set after this number of steps")
tf.flags.DEFINE_integer("checkpoint_every", 50, "Evaluate model on dev set after this number of steps")
# Session Parameters
tf.flags.DEFINE_boolean("allow_soft_placement", False, "Allow soft device placement (e.g. no GPU)")
tf.flags.DEFINE_boolean("log_device_placement", False, "Log placement of ops on devices")
# Print parameters
FLAGS = tf.flags.FLAGS
