def __init__(self,
vocab: Vocabulary,
source_embedder: TextFieldEmbedder,
encoder: Seq2SeqEncoder,
max_decoding_steps: int,
attention: Attention = None,
attention_function: SimilarityFunction = None,
beam_size: int = None,
target_namespace: str = "tokens",
target_embedding_dim: int = None,
scheduled_sampling_ratio: float = 0.,
use_bleu: bool = True) -> None:
super(Seq2Seq,
self).__init__(vocab,
source_embedder,
encoder,
max_decoding_steps,
attention=attention,
attention_function=attention_function,
beam_size=beam_size,
target_namespace=target_namespace,
target_embedding_dim=target_embedding_dim,
scheduled_sampling_ratio=scheduled_sampling_ratio,
use_bleu=use_bleu)
self._seqacc = SequenceAccuracy()
self._train_seqacc = SequenceAccuracy()
def test_sequence_accuracy(self):
accuracy = SequenceAccuracy()
gold = torch.Tensor([[1, 2, 3], [2, 4, 8], [0, 1, 1]])
predictions = torch.Tensor([[[1, 2, 3], [1, 2, -1]],
[[2, 4, 8], [2, 5, 9]],
[[-1, -1, -1], [0, 1, -1]]])
accuracy(predictions, gold)
actual_accuracy = accuracy.get_metric()
numpy.testing.assert_almost_equal(actual_accuracy, 2 / 3)
def test_sequence_accuracy_accumulates_and_resets_correctly(self):
accuracy = SequenceAccuracy()
gold = torch.Tensor([[1, 2, 3]])
accuracy(torch.Tensor([[[1, 2, 3]]]), gold)
accuracy(torch.Tensor([[[1, 2, 4]]]), gold)
actual_accuracy = accuracy.get_metric(reset=True)
numpy.testing.assert_almost_equal(actual_accuracy, 1 / 2)
assert accuracy.correct_count == 0
assert accuracy.total_count == 0
def test_sequence_accuracy_accumulates_and_resets_correctly(self):
accuracy = SequenceAccuracy()
gold = torch.Tensor([[1, 2, 3]])
accuracy(torch.Tensor([[[1, 2, 3]]]), gold)
accuracy(torch.Tensor([[[1, 2, 4]]]), gold)
actual_accuracy = accuracy.get_metric(reset=True)
numpy.testing.assert_almost_equal(actual_accuracy, 1/2)
assert accuracy.correct_count == 0
assert accuracy.total_count == 0
def test_sequence_accuracy_accumulates_and_resets_correctly(self, device: str):
accuracy = SequenceAccuracy()
gold = torch.tensor([[1, 2, 3]], device=device)
accuracy(torch.tensor([[[1, 2, 3]]], device=device), gold)
accuracy(torch.tensor([[[1, 2, 4]]], device=device), gold)
actual_accuracy = accuracy.get_metric(reset=True)
assert_allclose(actual_accuracy, 1 / 2)
assert accuracy.correct_count == 0
assert accuracy.total_count == 0
def test_sequence_accuracy(self, device: str):
accuracy = SequenceAccuracy()
gold = torch.tensor([[1, 2, 3], [2, 4, 8], [0, 1, 1]], device=device)
predictions = torch.tensor(
[[[1, 2, 3], [1, 2, -1]], [[2, 4, 8], [2, 5, 9]], [[-1, -1, -1], [0, 1, -1]]],
device=device,
)
accuracy(predictions, gold)
actual_accuracy = accuracy.get_metric()
assert_allclose(actual_accuracy, 2 / 3)
def test_sequence_accuracy_respects_mask(self):
accuracy = SequenceAccuracy()
gold = torch.Tensor([[1, 2, 3], [2, 4, 8], [0, 1, 1], [11, 13, 17]])
predictions = torch.Tensor([
[[1, 2, 3], [1, 2, -1]],
[[2, 4, 8], [2, 5, 9]],
[[-1, -1, -1], [0, 1, -1]],
[[12, 13, 17], [11, 13, 18]],
])
mask = torch.Tensor([[0, 1, 1], [1, 1, 1], [1, 1, 0], [1, 0, 1]])
accuracy(predictions, gold, mask)
actual_accuracy = accuracy.get_metric()
numpy.testing.assert_almost_equal(actual_accuracy, 3 / 4)
def test_sequence_accuracy(self):
accuracy = SequenceAccuracy()
gold = torch.Tensor([
[1, 2, 3],
[2, 4, 8],
[0, 1, 1]
])
predictions = torch.Tensor([
[[1, 2, 3], [1, 2, -1]],
[[2, 4, 8], [2, 5, 9]],
[[-1, -1, -1], [0, 1, -1]]
])
accuracy(predictions, gold)
actual_accuracy = accuracy.get_metric()
numpy.testing.assert_almost_equal(actual_accuracy, 2/3)
def test_distributed_sequence_accuracy(self):
gold = torch.tensor([[1, 2, 3], [2, 4, 8], [0, 1, 1], [11, 13, 17]])
predictions = torch.tensor([
[[1, 2, 3], [1, 2, -1]],
[[2, 4, 8], [2, 5, 9]],
[[-1, -1, -1], [0, 1, -1]],
[[12, 13, 17], [11, 13, 18]],
])
mask = torch.tensor([[False, True, True], [True, True, True],
[True, True, False], [True, False, True]], )
gold = [gold[:2], gold[2:]]
predictions = [predictions[:2], predictions[2:]]
mask = [mask[:2], mask[2:]]
metric_kwargs = {
"predictions": predictions,
"gold_labels": gold,
"mask": mask
}
desired_values = {"accuracy": 3 / 4}
run_distributed_test(
[-1, -1],
global_distributed_metric,
SequenceAccuracy(),
metric_kwargs,
desired_values,
exact=False,
)
def __init__(
self,
vocab: Vocabulary,
source_embedder: TextFieldEmbedder,
transformer: Dict,
max_decoding_steps: int,
target_namespace: str,
target_embedder: TextFieldEmbedder = None,
use_bleu: bool = True,
) -> None:
super().__init__(vocab)
self._target_namespace = target_namespace
self._start_index = self.vocab.get_token_index(START_SYMBOL,
self._target_namespace)
self._end_index = self.vocab.get_token_index(END_SYMBOL,
self._target_namespace)
self._pad_index = self.vocab.get_token_index(self.vocab._padding_token,
self._target_namespace)
if use_bleu:
self._bleu = BLEU(exclude_indices={
self._pad_index, self._end_index, self._start_index
})
else:
self._bleu = None
self._seq_acc = SequenceAccuracy()
self._max_decoding_steps = max_decoding_steps
self._source_embedder = source_embedder
self._ndim = transformer["d_model"]
self.pos_encoder = PositionalEncoding(self._ndim,
transformer["dropout"])
num_classes = self.vocab.get_vocab_size(self._target_namespace)
self._transformer = Transformer(**transformer)
self._transformer.apply(inplace_relu)
if target_embedder is None:
self._target_embedder = self._source_embedder
else:
self._target_embedder = target_embedder
self._output_projection_layer = Linear(self._ndim, num_classes)
def test_sequence_accuracy_respects_mask(self, device: str):
accuracy = SequenceAccuracy()
gold = torch.tensor([[1, 2, 3], [2, 4, 8], [0, 1, 1], [11, 13, 17]],
device=device)
predictions = torch.tensor(
[
[[1, 2, 3], [1, 2, -1]],
[[2, 4, 8], [2, 5, 9]],
[[-1, -1, -1], [0, 1, -1]],
[[12, 13, 17], [11, 13, 18]],
],
device=device,
)
mask = torch.tensor([[0, 1, 1], [1, 1, 1], [1, 1, 0], [1, 0, 1]],
device=device)
accuracy(predictions, gold, mask)
actual_accuracy = accuracy.get_metric()
assert_allclose(actual_accuracy, 3 / 4)
def test_sequence_accuracy_respects_mask(self):
accuracy = SequenceAccuracy()
gold = torch.Tensor([
[1, 2, 3],
[2, 4, 8],
[0, 1, 1],
[11, 13, 17],
])
predictions = torch.Tensor([
[[1, 2, 3], [1, 2, -1]],
[[2, 4, 8], [2, 5, 9]],
[[-1, -1, -1], [0, 1, -1]],
[[12, 13, 17], [11, 13, 18]]
])
mask = torch.Tensor([
[0, 1, 1],
[1, 1, 1],
[1, 1, 0],
[1, 0, 1]
])
accuracy(predictions, gold, mask)
actual_accuracy = accuracy.get_metric()
numpy.testing.assert_almost_equal(actual_accuracy, 3/4)
def multiple_runs(
global_rank: int,
world_size: int,
gpu_id: Union[int, torch.device],
metric: SequenceAccuracy,
metric_kwargs: Dict[str, List[Any]],
desired_values: Dict[str, Any],
exact: Union[bool, Tuple[float, float]] = True,
):
kwargs = {}
# Use the arguments meant for the process with rank `global_rank`.
for argname in metric_kwargs:
kwargs[argname] = metric_kwargs[argname][global_rank]
for i in range(200):
metric(**kwargs)
assert desired_values["accuracy"] == metric.get_metric()["accuracy"]
def __init__(
self,
vocab: Vocabulary,
source_embedder: TextFieldEmbedder,
transformer: Dict,
max_decoding_steps: int,
target_embedders: Dict[str, TextFieldEmbedder] = None,
loss_coefs: Dict = None,
) -> None:
super().__init__(vocab)
self._target_namespaces = loss_coefs.keys()
self._decoder_namespaces = transformer.get("num_decoder_layers", {}).keys()
self._start_index_dict = {k: self.vocab.get_token_index(START_SYMBOL, k) for k in self._decoder_namespaces}
self._end_index_dict = {k: self.vocab.get_token_index(END_SYMBOL, k) for k in self._decoder_namespaces}
self._pad_index_dict = {k: self.vocab.get_token_index(self.vocab._padding_token, k) for k in
self._target_namespaces}
self._loss_coefs = loss_coefs
self._metrics = {}
for tn in self._target_namespaces:
self._metrics[f'{tn}_acc'] = SequenceAccuracy()
self._max_decoding_steps = max_decoding_steps
self._source_embedder = source_embedder
self._ndim = transformer["d_model"]
self.pos_encoder = PositionalEncoding(self._ndim, transformer["dropout"])
self._transformer = MultiDecodersTPTransformer(**transformer)
self._transformer.apply(inplace_relu)
self._target_embedders = ModuleDict(target_embedders.items())
output_projection_layers = {}
for tn in self._target_namespaces:
num_classes = self.vocab.get_vocab_size(tn)
output_projection_layers[tn] = Linear(self._ndim, num_classes)
self._output_projection_layers = ModuleDict(output_projection_layers.items())
# tp parameters
self._embedding_project_layer = Linear(self._ndim, self._ndim)
def test_get_metric_on_new_object_works(self, device: str):
accuracy = SequenceAccuracy()
actual_accuracy = accuracy.get_metric(reset=True)["accuracy"]
assert_allclose(actual_accuracy, 0)
class Seq2Seq(SimpleSeq2Seq):
def __init__(self,
vocab: Vocabulary,
source_embedder: TextFieldEmbedder,
encoder: Seq2SeqEncoder,
max_decoding_steps: int,
attention: Attention = None,
attention_function: SimilarityFunction = None,
beam_size: int = None,
target_namespace: str = "tokens",
target_embedding_dim: int = None,
scheduled_sampling_ratio: float = 0.,
use_bleu: bool = True) -> None:
super(Seq2Seq, self).__init__(vocab, source_embedder,encoder,max_decoding_steps,
attention=attention, attention_function=attention_function,
beam_size=beam_size,target_namespace=target_namespace,
target_embedding_dim=target_embedding_dim,
scheduled_sampling_ratio=scheduled_sampling_ratio,
use_bleu=use_bleu)
self._seqacc = SequenceAccuracy()
@overrides
def forward(self, # type: ignore
nl: Dict[str, torch.LongTensor],
fl: Dict[str, torch.LongTensor] = None, id=None) -> Dict[str, torch.Tensor]:
# pylint: disable=arguments-differ
"""
Make foward pass with decoder logic for producing the entire target sequence.
Parameters
----------
source_tokens : ``Dict[str, torch.LongTensor]``
The output of `TextField.as_array()` applied on the source `TextField`. This will be
passed through a `TextFieldEmbedder` and then through an encoder.
target_tokens : ``Dict[str, torch.LongTensor]``, optional (default = None)
Output of `Textfield.as_array()` applied on target `TextField`. We assume that the
target tokens are also represented as a `TextField`.
Returns
-------
Dict[str, torch.Tensor]
"""
source_tokens, target_tokens = nl, fl
state = self._encode(source_tokens)
if target_tokens:
state = self._init_decoder_state(state)
# The `_forward_loop` decodes the input sequence and computes the loss during training
# and validation.
output_dict = self._forward_loop(state, target_tokens)
else:
output_dict = {}
if True: #not self.training:
state = self._init_decoder_state(state)
predictions = self._forward_beam_search(state)
output_dict.update(predictions)
# shape: (batch_size, beam_size, max_sequence_length)
top_k_predictions = output_dict["predictions"]
# shape: (batch_size, max_predicted_sequence_length)
best_predictions = top_k_predictions[:, 0, :]
if target_tokens and self._bleu:
self._bleu(best_predictions, target_tokens["tokens"])
if target_tokens:
seqacc_gold = target_tokens["tokens"][:, 1:]
self._seqacc(best_predictions.unsqueeze(1)[:, :, :seqacc_gold.size(1)],
seqacc_gold,
mask=(seqacc_gold != 0).long())
return output_dict
@overrides
def get_metrics(self, reset: bool = False) -> Dict[str, float]:
all_metrics: Dict[str, float] = {}
if self._bleu:
all_metrics.update(self._bleu.get_metric(reset=reset))
all_metrics.update({"SeqAcc": self._seqacc.get_metric(reset=reset)})
return all_metrics
def test_get_metric_on_new_object_works(self):
accuracy = SequenceAccuracy()
actual_accuracy = accuracy.get_metric(reset=True)
numpy.testing.assert_almost_equal(actual_accuracy, 0)
def test_get_metric_on_new_object_works(self):
accuracy = SequenceAccuracy()
actual_accuracy = accuracy.get_metric(reset=True)
numpy.testing.assert_almost_equal(actual_accuracy, 0)
class MyTransformer(Model):
def __init__(
self,
vocab: Vocabulary,
source_embedder: TextFieldEmbedder,
transformer: Dict,
max_decoding_steps: int,
target_namespace: str,
target_embedder: TextFieldEmbedder = None,
use_bleu: bool = True,
) -> None:
super().__init__(vocab)
self._target_namespace = target_namespace
self._start_index = self.vocab.get_token_index(START_SYMBOL,
self._target_namespace)
self._end_index = self.vocab.get_token_index(END_SYMBOL,
self._target_namespace)
self._pad_index = self.vocab.get_token_index(self.vocab._padding_token,
self._target_namespace)
if use_bleu:
self._bleu = BLEU(exclude_indices={
self._pad_index, self._end_index, self._start_index
})
else:
self._bleu = None
self._seq_acc = SequenceAccuracy()
self._max_decoding_steps = max_decoding_steps
self._source_embedder = source_embedder
self._ndim = transformer["d_model"]
self.pos_encoder = PositionalEncoding(self._ndim,
transformer["dropout"])
num_classes = self.vocab.get_vocab_size(self._target_namespace)
self._transformer = Transformer(**transformer)
self._transformer.apply(inplace_relu)
if target_embedder is None:
self._target_embedder = self._source_embedder
else:
self._target_embedder = target_embedder
self._output_projection_layer = Linear(self._ndim, num_classes)
def _get_mask(self, meta_data):
mask = torch.zeros(1, len(meta_data),
self.vocab.get_vocab_size(
self._target_namespace)).float()
for bidx, md in enumerate(meta_data):
for k, v in self.vocab._token_to_index[
self._target_namespace].items():
if 'position' in k and k not in md['avail_pos']:
mask[:, bidx, v] = float('-inf')
return mask
def generate_square_subsequent_mask(self, sz):
mask = (torch.triu(torch.ones(sz, sz)) == 1).transpose(0, 1)
mask = mask.float().masked_fill(mask == False,
float('-inf')).masked_fill(
mask == True, float(0.0))
return mask
@overrides
def forward(
self,
source_tokens: Dict[str, torch.LongTensor],
target_tokens: Dict[str, torch.LongTensor] = None,
meta_data: Any = None,
) -> Dict[str, torch.Tensor]:
src, src_key_padding_mask = self._encode(self._source_embedder,
source_tokens)
memory = self._transformer.encoder(
src, src_key_padding_mask=src_key_padding_mask)
if meta_data is not None:
target_vocab_mask = self._get_mask(meta_data)
target_vocab_mask = target_vocab_mask.to(memory.device)
else:
target_vocab_mask = None
output_dict = {}
targets = None
if target_tokens:
targets = target_tokens["tokens"][:, 1:]
target_mask = (util.get_text_field_mask({"tokens": targets}) == 1)
assert targets.size(1) <= self._max_decoding_steps
if self.training and target_tokens:
tgt, tgt_key_padding_mask = self._encode(
self._target_embedder,
{"tokens": target_tokens["tokens"][:, :-1]})
tgt_mask = self.generate_square_subsequent_mask(tgt.size(0)).to(
memory.device)
output = self._transformer.decoder(
tgt,
memory,
tgt_mask=tgt_mask,
tgt_key_padding_mask=tgt_key_padding_mask,
memory_key_padding_mask=src_key_padding_mask)
logits = self._output_projection_layer(output)
if target_vocab_mask is not None:
logits += target_vocab_mask
class_probabilities = F.softmax(logits.detach(), dim=-1)
_, predictions = torch.max(class_probabilities, -1)
logits = logits.transpose(0, 1)
loss = self._get_loss(logits, targets, target_mask)
output_dict["loss"] = loss
else:
assert self.training is False
output_dict["loss"] = torch.tensor(0.0).to(memory.device)
if targets is not None:
max_target_len = targets.size(1)
else:
max_target_len = None
predictions, class_probabilities = self._decoder_step_by_step(
memory,
src_key_padding_mask,
target_vocab_mask,
max_target_len=max_target_len)
predictions = predictions.transpose(0, 1)
output_dict["predictions"] = predictions
output_dict["class_probabilities"] = class_probabilities.transpose(
0, 1)
if target_tokens:
with torch.no_grad():
best_predictions = output_dict["predictions"]
if self._bleu:
self._bleu(best_predictions, targets)
batch_size = targets.size(0)
max_sz = max(best_predictions.size(1), targets.size(1),
target_mask.size(1))
best_predictions_ = torch.zeros(batch_size,
max_sz).to(memory.device)
best_predictions_[:, :best_predictions.
size(1)] = best_predictions
targets_ = torch.zeros(batch_size, max_sz).to(memory.device)
targets_[:, :targets.size(1)] = targets.cpu()
target_mask_ = torch.zeros(batch_size,
max_sz).to(memory.device)
target_mask_[:, :target_mask.size(1)] = target_mask
self._seq_acc(best_predictions_.unsqueeze(1), targets_,
target_mask_)
return output_dict
@overrides
def decode(
self, output_dict: Dict[str,
torch.Tensor]) -> Dict[str, torch.Tensor]:
predicted_indices = output_dict["predictions"]
if not isinstance(predicted_indices, numpy.ndarray):
# shape: (batch_size, num_decoding_steps)
predicted_indices = predicted_indices.detach().cpu().numpy()
# class_probabilities = output_dict["class_probabilities"].detach().cpu()
# sample_predicted_indices = []
# for cp in class_probabilities:
# sample = torch.multinomial(cp, num_samples=1)
# sample_predicted_indices.append(sample)
# # shape: (batch_size, num_decoding_steps, num_samples)
# sample_predicted_indices = torch.stack(sample_predicted_indices)
all_predicted_tokens = []
for indices in predicted_indices:
# Beam search gives us the top k results for each source sentence in the batch
# but we just want the single best.
if len(indices.shape) > 1:
indices = indices[0]
indices = list(indices)
# Collect indices till the first end_symbol
if self._end_index in indices:
indices = indices[:indices.index(self._end_index)]
predicted_tokens = [
self.vocab.get_token_from_index(
x, namespace=self._target_namespace) for x in indices
]
all_predicted_tokens.append(predicted_tokens)
output_dict["predicted_tokens"] = all_predicted_tokens
return output_dict
def _encode(
self, embedder: TextFieldEmbedder,
tokens: Dict[str,
torch.Tensor]) -> Tuple[torch.Tensor, torch.Tensor]:
src = embedder(tokens) * math.sqrt(self._ndim)
src = src.transpose(0, 1)
src = self.pos_encoder(src)
mask = util.get_text_field_mask(tokens)
mask = (mask == 0)
return src, mask
def _decoder_step_by_step(
self,
memory: torch.Tensor,
memory_key_padding_mask: torch.Tensor,
target_vocab_mask: torch.Tensor = None,
max_target_len: int = None) -> Tuple[torch.Tensor, torch.Tensor]:
batch_size = memory.size(1)
if getattr(self, "target_limit_decode_steps",
False) and max_target_len is not None:
num_decoding_steps = min(self._max_decoding_steps, max_target_len)
print('decoding steps: ', num_decoding_steps)
else:
num_decoding_steps = self._max_decoding_steps
last_predictions = memory.new_full(
(batch_size, ), fill_value=self._start_index).long()
step_predictions: List[torch.Tensor] = []
all_predicts = memory.new_full((batch_size, num_decoding_steps),
fill_value=0).long()
for timestep in range(num_decoding_steps):
all_predicts[:, timestep] = last_predictions
tgt, tgt_key_padding_mask = self._encode(
self._target_embedder,
{"tokens": all_predicts[:, :timestep + 1]})
tgt_mask = self.generate_square_subsequent_mask(timestep + 1).to(
memory.device)
output = self._transformer.decoder(
tgt,
memory,
tgt_mask=tgt_mask,
tgt_key_padding_mask=tgt_key_padding_mask,
memory_key_padding_mask=memory_key_padding_mask)
output_projections = self._output_projection_layer(output)
if target_vocab_mask is not None:
output_projections += target_vocab_mask
class_probabilities = F.softmax(output_projections, dim=-1)
_, predicted_classes = torch.max(class_probabilities, -1)
# shape (predicted_classes): (batch_size,)
last_predictions = predicted_classes[timestep, :]
step_predictions.append(last_predictions)
if ((last_predictions == self._end_index) +
(last_predictions == self._pad_index)).all():
break
# shape: (num_decoding_steps, batch_size)
predictions = torch.stack(step_predictions)
return predictions, class_probabilities
@staticmethod
def _get_loss(logits: torch.FloatTensor, targets: torch.LongTensor,
target_mask: torch.FloatTensor) -> torch.Tensor:
logits = logits.contiguous()
# shape: (batch_size, num_decoding_steps)
relevant_targets = targets.contiguous()
# shape: (batch_size, num_decoding_steps)
relevant_mask = target_mask.contiguous()
return util.sequence_cross_entropy_with_logits(logits,
relevant_targets,
relevant_mask)
@overrides
def get_metrics(self, reset: bool = False) -> Dict[str, float]:
all_metrics: Dict[str, float] = {}
if self._bleu:
all_metrics.update(self._bleu.get_metric(reset=reset))
all_metrics['seq_acc'] = self._seq_acc.get_metric(reset=reset)
return all_metrics
def load_state_dict(self, state_dict, strict=True):
new_state_dict = {}
for k, v in state_dict.items():
if k.startswith('module.'):
new_state_dict[k[len('module.'):]] = v
else:
new_state_dict[k] = v
super(MyTransformer, self).load_state_dict(new_state_dict, strict)
