def test_trainable_variables(self):
"""Tests the functionality of automatically collecting trainable
variables.
"""
inputs = tf.placeholder(dtype=tf.int32, shape=[None, None])
# case 1
hparams = {
"pretrained_model_name": None,
}
clas = BERTClassifier(hparams=hparams)
_, _ = clas(inputs)
self.assertEqual(len(clas.trainable_variables), 199 + 2)
# case 2
hparams = {
"pretrained_model_name": None,
"clas_strategy": "all_time",
"max_seq_length": 8,
}
clas = BERTClassifier(hparams=hparams)
_, _ = clas(inputs)
self.assertEqual(len(clas.trainable_variables), 199 + 2)
# case 2
hparams = {
"pretrained_model_name": None,
"clas_strategy": "time_wise",
}
clas = BERTClassifier(hparams=hparams)
_, _ = clas(inputs)
self.assertEqual(len(clas.trainable_variables), 199 + 2)
def test_model_loading(self):
r"""Tests model loading functionality."""
inputs = tf.placeholder(dtype=tf.int32, shape=[None, None])
for pretrained_model_name in BERTClassifier.available_checkpoints():
classifier = BERTClassifier(
pretrained_model_name=pretrained_model_name)
_, _ = classifier(inputs)
def test_encode(self):
"""Tests encoding.
"""
max_time = 8
batch_size = 16
inputs = tf.random_uniform([batch_size, max_time],
maxval=30521,
dtype=tf.int32)
# case 1
hparams = {
"pretrained_model_name": None,
}
clas = BERTClassifier(hparams=hparams)
logits, pred = clas(inputs)
with self.test_session() as sess:
sess.run(tf.global_variables_initializer())
logits_, pred_ = sess.run([logits, pred])
self.assertEqual(logits_.shape,
(batch_size, clas.hparams.num_classes))
self.assertEqual(pred_.shape, (batch_size, ))
# case 2
hparams = {
"pretrained_model_name": None,
"num_classes": 10,
"clas_strategy": "time_wise"
}
clas = BERTClassifier(hparams=hparams)
logits, pred = clas(inputs)
with self.test_session() as sess:
sess.run(tf.global_variables_initializer())
logits_, pred_ = sess.run([logits, pred])
self.assertEqual(logits_.shape,
(batch_size, max_time, clas.hparams.num_classes))
self.assertEqual(pred_.shape, (batch_size, max_time))
# case 3
hparams = {
"pretrained_model_name": None,
"num_classes": 0,
"clas_strategy": "time_wise"
}
clas = BERTClassifier(hparams=hparams)
logits, pred = clas(inputs)
with self.test_session() as sess:
sess.run(tf.global_variables_initializer())
logits_, pred_ = sess.run([logits, pred])
self.assertEqual(logits_.shape,
(batch_size, max_time, clas.hparams.encoder.dim))
self.assertEqual(pred_.shape, (batch_size, max_time))
# case 4
hparams = {
"pretrained_model_name": None,
"num_classes": 10,
"clas_strategy": "all_time",
"max_seq_length": max_time
}
inputs = tf.placeholder(tf.int32, shape=[batch_size, 6])
clas = BERTClassifier(hparams=hparams)
logits, pred = clas(inputs)
with self.test_session() as sess:
sess.run(tf.global_variables_initializer())
logits_, pred_ = sess.run([logits, pred],
feed_dict={
inputs:
np.random.randint(30521,
size=(batch_size,
6))
})
self.assertEqual(logits_.shape,
(batch_size, clas.hparams.num_classes))
self.assertEqual(pred_.shape, (batch_size, ))