def test_categorical_embedding_to_encoding_bilstm_sequence_of_words(self):
"""
Check that categorical_embedding_to_encoding_bilstm function does not fail with sequences of word (in a single
row).
"""
embedding_size = 32
encoding_size = 64
feature_info = {
"name": "categorical_variable",
"feature_layer_info": {
"type": "numeric",
"fn": "categorical_embedding_to_encoding_bilstm",
"args": {
"vocabulary_file":
"ml4ir/applications/classification/tests/data/configs/vocabulary/entity_id.csv",
"encoding_type": "bilstm",
"encoding_size": encoding_size,
"embedding_size": embedding_size,
"dropout_rate": 0.2,
},
},
"tfrecord_type": SequenceExampleTypeKey.CONTEXT,
}
string_tensor = tf.constant([[["AAA", "BBB", "out_of_vocabulary"]]])
categorical_fns.categorical_embedding_to_encoding_bilstm(
string_tensor, feature_info, self.file_io)
def test_categorical_embedding_to_encoding_bilstm_file_truncation(self):
"""
Asserts the conversion of a string tensor to its corresponding sequence encoding
obtained through the categorical_embedding_to_encoding_bilstm function
Works by converting each string into a bytes sequence and then
passing it through a biLSTM.
The embedding and encoding dimensions are controlled by the feature_info
"""
max_length = 2
embedding_size = 128
encoding_size = 512
feature_info = {
"name": "categorical_variable",
"feature_layer_info": {
"type": "numeric",
"fn": "categorical_embedding_to_encoding_bilstm",
"args": {
"vocabulary_file":
"ml4ir/applications/classification/tests/data/configs/vocabulary/entity_id.csv",
"encoding_type": "bilstm",
"encoding_size": encoding_size,
"embedding_size": embedding_size,
"max_length": max_length,
},
},
"tfrecord_type": SequenceExampleTypeKey.CONTEXT,
}
# Define an input string tensor
string_tensor = tf.constant([[["AAA"]], [["BBB"]], [["AAA"]],
[["CCC"]], [["out_of_vocabulary"]]])
sequence_encoding = categorical_fns.categorical_embedding_to_encoding_bilstm(
string_tensor, feature_info, self.file_io)
# Assert the right shapes of the resulting encoding based on the feature_info
assert sequence_encoding.shape[0] == len(string_tensor)
assert sequence_encoding.shape[1] == 1
assert sequence_encoding.shape[2] == encoding_size
assert tf.reduce_all(
tf.equal(sequence_encoding[0], sequence_encoding[2]))
assert not tf.reduce_all(
tf.equal(sequence_encoding[0], sequence_encoding[1]))
assert tf.reduce_all(
tf.equal(sequence_encoding[3], sequence_encoding[4]))
assert not tf.reduce_all(
tf.equal(sequence_encoding[1], sequence_encoding[4]))
def test_categorical_embedding_to_encoding_bilstm_oov_mapping_with_dropout(
self):
"""
Asserts "out of vocabulary" words are mapped to the same value when dropout_rate is used,
while word "in vocabulary" are mapped to their own embeddings.
"""
embedding_size = 128
encoding_size = 512
vocabulary_file = (
"ml4ir/applications/classification/tests/data/configs/vocabulary/entity_id.csv"
)
feature_info = {
"name": "categorical_variable",
"feature_layer_info": {
"type": "numeric",
"fn": "categorical_embedding_to_encoding_bilstm",
"args": {
"vocabulary_file": vocabulary_file,
"encoding_type": "bilstm",
"encoding_size": encoding_size,
"embedding_size": embedding_size,
"dropout_rate": 0.2,
},
},
"tfrecord_type": SequenceExampleTypeKey.CONTEXT,
}
df_vocabulary = self.file_io.read_df(vocabulary_file)
vocabulary = df_vocabulary.values.flatten().tolist()
# Define a batch where each row is a single word from the vocabulary (so that we can later compare each word
# encoding)
words = vocabulary + ["out_of_vocabulary", "out_of_vocabulary2"]
string_tensor = tf.constant([[[word]] for word in words])
# If the embedding lookup input dimension were computed incorrectly, the following call would fail with:
# tensorflow.python.framework.errors_impl.InvalidArgumentError: indices[0,0,7] = 8 is not in [0, 8)
sequence_encoding = categorical_fns.categorical_embedding_to_encoding_bilstm(
string_tensor, feature_info, self.file_io)
# Strings 1 and 2 should result in the same embedding because they are both OOV
# Check that each word, except the OOV ones are mapped to different encodings (actually different embeddings,
# but the encoding on top should be equivalent as each row is a single word).
for word1_position in range(0, len(vocabulary)):
# +1 to include one OOV word
for word2_position in range(word1_position + 1,
len(vocabulary) + 1):
if tf.reduce_all(
tf.equal(sequence_encoding[word1_position],
sequence_encoding[word2_position])):
word1, word2 = words[word1_position], words[word2_position]
self.assertFalse(
tf.reduce_all(
tf.equal(
sequence_encoding[word1_position],
sequence_encoding[word2_position],
)),
msg=
"Check that non-OOV words map to different encodings, and not also not to OOV:"
"{} vs. {}".format(word1, word2),
)
self.assertTrue(
tf.reduce_all(
tf.equal(sequence_encoding[-1], sequence_encoding[-2])),
msg="Check that OOV words map to the same encodings",
)
assert not tf.reduce_all(
tf.equal(sequence_encoding[0], sequence_encoding[1]))