def __init__(self,
vocab: Vocabulary,
question_embedder: TextFieldEmbedder,
action_embedding_dim: int,
encoder: Seq2SeqEncoder,
entity_encoder: Seq2VecEncoder,
decoder_beam_search: BeamSearch,
max_decoding_steps: int,
attention: Attention,
mixture_feedforward: FeedForward = None,
add_action_bias: bool = True,
training_beam_size: int = None,
use_neighbor_similarity_for_linking: bool = False,
dropout: float = 0.0,
num_linking_features: int = 10,
rule_namespace: str = 'rule_labels',
tables_directory: str = '/wikitables/') -> None:
use_similarity = use_neighbor_similarity_for_linking
super().__init__(vocab=vocab,
question_embedder=question_embedder,
action_embedding_dim=action_embedding_dim,
encoder=encoder,
entity_encoder=entity_encoder,
max_decoding_steps=max_decoding_steps,
add_action_bias=add_action_bias,
use_neighbor_similarity_for_linking=use_similarity,
dropout=dropout,
num_linking_features=num_linking_features,
rule_namespace=rule_namespace,
tables_directory=tables_directory)
self._beam_search = decoder_beam_search
self._decoder_trainer = MaximumMarginalLikelihood(training_beam_size)
self._decoder_step = LinkingTransitionFunction(
encoder_output_dim=self._encoder.get_output_dim(),
action_embedding_dim=action_embedding_dim,
input_attention=attention,
num_start_types=self._num_start_types,
predict_start_type_separately=True,
add_action_bias=self._add_action_bias,
mixture_feedforward=mixture_feedforward,
dropout=dropout)
def __init__(
self,
vocab: Vocabulary,
question_embedder: TextFieldEmbedder,
action_embedding_dim: int,
encoder: Seq2SeqEncoder,
decoder_beam_search: BeamSearch,
max_decoding_steps: int,
attention: Attention,
mixture_feedforward: FeedForward = None,
add_action_bias: bool = True,
dropout: float = 0.0,
num_linking_features: int = 0,
num_entity_bits: int = 0,
entity_bits_output: bool = True,
use_entities: bool = False,
denotation_only: bool = False,
# Deprecated parameter to load older models
entity_encoder: Seq2VecEncoder = None, # pylint: disable=unused-argument
entity_similarity_mode: str = "dot_product",
rule_namespace: str = 'rule_labels') -> None:
super(QuarelSemanticParser, self).__init__(vocab)
self._question_embedder = question_embedder
self._encoder = encoder
self._beam_search = decoder_beam_search
self._max_decoding_steps = max_decoding_steps
if dropout > 0:
self._dropout = torch.nn.Dropout(p=dropout)
else:
self._dropout = lambda x: x
self._rule_namespace = rule_namespace
self._denotation_accuracy = Average()
self._action_sequence_accuracy = Average()
self._has_logical_form = Average()
self._embedding_dim = question_embedder.get_output_dim()
self._use_entities = use_entities
# Note: there's only one non-trivial entity type in QuaRel for now, so most of the
# entity_type stuff is irrelevant
self._num_entity_types = 4 # TODO(mattg): get this in a more principled way somehow?
self._num_start_types = 1 # Hardcoded until we feed lf syntax into the model
self._entity_type_encoder_embedding = Embedding(
self._num_entity_types, self._embedding_dim)
self._entity_type_decoder_embedding = Embedding(
self._num_entity_types, action_embedding_dim)
self._entity_similarity_layer = None
self._entity_similarity_mode = entity_similarity_mode
if self._entity_similarity_mode == "weighted_dot_product":
self._entity_similarity_layer = \
TimeDistributed(torch.nn.Linear(self._embedding_dim, 1, bias=False))
# Center initial values around unweighted dot product
self._entity_similarity_layer._module.weight.data += 1 # pylint: disable=protected-access
elif self._entity_similarity_mode == "dot_product":
pass
else:
raise ValueError("Invalid entity_similarity_mode: {}".format(
self._entity_similarity_mode))
if num_linking_features > 0:
self._linking_params = torch.nn.Linear(num_linking_features, 1)
else:
self._linking_params = None
self._decoder_trainer = MaximumMarginalLikelihood()
self._encoder_output_dim = self._encoder.get_output_dim()
if entity_bits_output:
self._encoder_output_dim += num_entity_bits
self._entity_bits_output = entity_bits_output
self._debug_count = 10
self._num_denotation_cats = 2 # Hardcoded for simplicity
self._denotation_only = denotation_only
if self._denotation_only:
self._denotation_accuracy_cat = CategoricalAccuracy()
self._denotation_classifier = torch.nn.Linear(
self._encoder_output_dim, self._num_denotation_cats)
# Rest of init not needed for denotation only where no decoding to actions needed
return
self._action_padding_index = -1 # the padding value used by IndexField
num_actions = vocab.get_vocab_size(self._rule_namespace)
self._num_actions = num_actions
self._action_embedder = Embedding(num_embeddings=num_actions,
embedding_dim=action_embedding_dim)
# We are tying the action embeddings used for input and output
# self._output_action_embedder = Embedding(num_embeddings=num_actions, embedding_dim=action_embedding_dim)
self._output_action_embedder = self._action_embedder # tied weights
self._add_action_bias = add_action_bias
if self._add_action_bias:
self._action_biases = Embedding(num_embeddings=num_actions,
embedding_dim=1)
# This is what we pass as input in the first step of decoding, when we don't have a
# previous action, or a previous question attention.
self._first_action_embedding = torch.nn.Parameter(
torch.FloatTensor(action_embedding_dim))
self._first_attended_question = torch.nn.Parameter(
torch.FloatTensor(self._encoder_output_dim))
torch.nn.init.normal_(self._first_action_embedding)
torch.nn.init.normal_(self._first_attended_question)
self._decoder_step = LinkingTransitionFunction(
encoder_output_dim=self._encoder_output_dim,
action_embedding_dim=action_embedding_dim,
input_attention=attention,
num_start_types=self._num_start_types,
predict_start_type_separately=False,
add_action_bias=self._add_action_bias,
mixture_feedforward=mixture_feedforward,
dropout=dropout)
def __init__(self,
question_embedder: TextFieldEmbedder,
input_memory_embedder: TextFieldEmbedder,
output_memory_embedder: TextFieldEmbedder,
question_encoder: Seq2SeqEncoder,
input_memory_encoder: Seq2VecEncoder,
output_memory_encoder: Seq2VecEncoder,
decoder_beam_search: BeamSearch,
input_attention: Attention,
past_attention: Attention,
action_embedding_dim: int,
max_decoding_steps: int,
nhop: int,
decoding_nhop: int,
vocab: Vocabulary,
dataset_path: str = 'dataset',
parse_sql_on_decoding: bool = True,
training_beam_size: int = None,
add_action_bias: bool = True,
decoder_self_attend: bool = True,
decoder_num_layers: int = 1,
dropout: float = 0.0,
rule_namespace: str = 'rule_labels') -> None:
super().__init__(vocab)
self.question_embedder = question_embedder
self._input_mm_embedder = input_memory_embedder
self._output_mm_embedder = output_memory_embedder
self._question_encoder = question_encoder
self._input_mm_encoder = TimeDistributed(input_memory_encoder)
self._output_mm_encoder = TimeDistributed(output_memory_encoder)
self.parse_sql_on_decoding = parse_sql_on_decoding
self._self_attend = decoder_self_attend
self._max_decoding_steps = max_decoding_steps
self._add_action_bias = add_action_bias
self._rule_namespace = rule_namespace
num_actions = vocab.get_vocab_size(self._rule_namespace)
if self._add_action_bias:
input_action_dim = action_embedding_dim + 1
else:
input_action_dim = action_embedding_dim
self._action_embedder = Embedding(num_embeddings=num_actions,
embedding_dim=input_action_dim)
self._input_action_embedder = Embedding(
num_embeddings=num_actions, embedding_dim=action_embedding_dim)
self._output_action_embedder = Embedding(
num_embeddings=num_actions, embedding_dim=action_embedding_dim)
self._num_entity_types = 9
self._entity_type_decoder_input_embedding = Embedding(
self._num_entity_types, action_embedding_dim)
self._entity_type_decoder_output_embedding = Embedding(
self._num_entity_types, action_embedding_dim)
self._entity_type_encoder_embedding = Embedding(
self._num_entity_types,
(int)(question_encoder.get_output_dim() / 2))
self._decoder_num_layers = decoder_num_layers
self._action_embedding_dim = action_embedding_dim
self._ent2ent_ff = FeedForward(action_embedding_dim, 1,
action_embedding_dim,
Activation.by_name('relu')())
if dropout > 0:
self._dropout = torch.nn.Dropout(p=dropout)
else:
self._dropout = lambda x: x
self._first_action_embedding = torch.nn.Parameter(
torch.FloatTensor(action_embedding_dim))
self._first_attended_utterance = torch.nn.Parameter(
torch.FloatTensor(question_encoder.get_output_dim()))
torch.nn.init.normal_(self._first_action_embedding)
torch.nn.init.normal_(self._first_attended_utterance)
if self._self_attend:
self._transition_function = AttendPastSchemaItemsTransitionFunction(
encoder_output_dim=question_encoder.get_output_dim(),
action_embedding_dim=action_embedding_dim,
input_attention=input_attention,
past_attention=past_attention,
decoding_nhop=decoding_nhop,
predict_start_type_separately=False,
add_action_bias=self._add_action_bias,
dropout=dropout,
num_layers=self._decoder_num_layers)
else:
self._transition_function = LinkingTransitionFunction(
encoder_output_dim=question_encoder.get_output_dim(),
action_embedding_dim=action_embedding_dim,
input_attention=input_attention,
predict_start_type_separately=False,
add_action_bias=self._add_action_bias,
dropout=dropout,
num_layers=self._decoder_num_layers)
self._mm_attn = MemAttn(question_encoder.get_output_dim(), nhop)
self._beam_search = decoder_beam_search
self._decoder_trainer = MaximumMarginalLikelihood(training_beam_size)
self._action_padding_index = -1 # the padding value used by IndexField
self._exact_match = Average()
self._sql_evaluator_match = Average()
self._action_similarity = Average()
self._acc_single = Average()
self._acc_multi = Average()
self._beam_hit = Average()
# TODO: Remove hard-coded dirs
self._evaluate_func = partial(
evaluate,
db_dir=os.path.join(dataset_path, 'database'),
table=os.path.join(dataset_path, 'tables.json'),
check_valid=False)