2020-06-03 12:45:00 +00:00
|
|
|
import numpy
|
|
|
|
|
|
|
|
from thinc.api import chain, Maxout, LayerNorm, Softmax, Linear, zero_init, Model
|
2020-02-27 17:42:27 +00:00
|
|
|
|
|
|
|
|
2020-06-12 00:02:07 +00:00
|
|
|
def build_multi_task_model(tok2vec, maxout_pieces, token_vector_width, nO=None):
|
|
|
|
softmax = Softmax(nO=nO, nI=token_vector_width * 2)
|
2020-02-27 17:42:27 +00:00
|
|
|
model = chain(
|
|
|
|
tok2vec,
|
2020-06-20 12:15:04 +00:00
|
|
|
Maxout(
|
|
|
|
nO=token_vector_width * 2,
|
|
|
|
nI=token_vector_width,
|
|
|
|
nP=maxout_pieces,
|
|
|
|
dropout=0.0,
|
|
|
|
),
|
2020-02-27 17:42:27 +00:00
|
|
|
LayerNorm(token_vector_width * 2),
|
2020-06-12 00:02:07 +00:00
|
|
|
softmax,
|
2020-02-27 17:42:27 +00:00
|
|
|
)
|
2020-06-12 00:02:07 +00:00
|
|
|
model.set_ref("tok2vec", tok2vec)
|
|
|
|
model.set_ref("output_layer", softmax)
|
2020-02-27 17:42:27 +00:00
|
|
|
return model
|
|
|
|
|
|
|
|
|
2020-06-12 00:02:07 +00:00
|
|
|
def build_cloze_multi_task_model(vocab, tok2vec, maxout_pieces, nO=None):
|
|
|
|
# nO = vocab.vectors.data.shape[1]
|
2020-02-27 17:42:27 +00:00
|
|
|
output_layer = chain(
|
|
|
|
Maxout(
|
2020-06-20 12:15:04 +00:00
|
|
|
nO=nO,
|
|
|
|
nI=tok2vec.get_dim("nO"),
|
|
|
|
nP=maxout_pieces,
|
|
|
|
normalize=True,
|
|
|
|
dropout=0.0,
|
2020-02-27 17:42:27 +00:00
|
|
|
),
|
2020-06-12 00:02:07 +00:00
|
|
|
Linear(nO=nO, nI=nO, init_W=zero_init),
|
2020-02-27 17:42:27 +00:00
|
|
|
)
|
|
|
|
model = chain(tok2vec, output_layer)
|
|
|
|
model = build_masked_language_model(vocab, model)
|
2020-06-12 00:02:07 +00:00
|
|
|
model.set_ref("tok2vec", tok2vec)
|
|
|
|
model.set_ref("output_layer", output_layer)
|
2020-02-27 17:42:27 +00:00
|
|
|
return model
|
|
|
|
|
|
|
|
|
2020-06-03 12:45:00 +00:00
|
|
|
def build_masked_language_model(vocab, wrapped_model, mask_prob=0.15):
|
|
|
|
"""Convert a model into a BERT-style masked language model"""
|
|
|
|
|
|
|
|
random_words = _RandomWords(vocab)
|
|
|
|
|
|
|
|
def mlm_forward(model, docs, is_train):
|
|
|
|
mask, docs = _apply_mask(docs, random_words, mask_prob=mask_prob)
|
|
|
|
mask = model.ops.asarray(mask).reshape((mask.shape[0], 1))
|
2020-06-20 12:15:04 +00:00
|
|
|
output, backprop = model.get_ref("wrapped-model").begin_update(
|
|
|
|
docs
|
|
|
|
) # drop=drop
|
2020-06-03 12:45:00 +00:00
|
|
|
|
|
|
|
def mlm_backward(d_output):
|
|
|
|
d_output *= 1 - mask
|
|
|
|
return backprop(d_output)
|
|
|
|
|
|
|
|
return output, mlm_backward
|
|
|
|
|
|
|
|
mlm_model = Model("masked-language-model", mlm_forward, layers=[wrapped_model])
|
|
|
|
mlm_model.set_ref("wrapped-model", wrapped_model)
|
|
|
|
|
|
|
|
return mlm_model
|
|
|
|
|
|
|
|
|
|
|
|
class _RandomWords(object):
|
|
|
|
def __init__(self, vocab):
|
|
|
|
self.words = [lex.text for lex in vocab if lex.prob != 0.0]
|
|
|
|
self.probs = [lex.prob for lex in vocab if lex.prob != 0.0]
|
|
|
|
self.words = self.words[:10000]
|
|
|
|
self.probs = self.probs[:10000]
|
|
|
|
self.probs = numpy.exp(numpy.array(self.probs, dtype="f"))
|
|
|
|
self.probs /= self.probs.sum()
|
|
|
|
self._cache = []
|
|
|
|
|
|
|
|
def next(self):
|
|
|
|
if not self._cache:
|
|
|
|
self._cache.extend(
|
|
|
|
numpy.random.choice(len(self.words), 10000, p=self.probs)
|
|
|
|
)
|
|
|
|
index = self._cache.pop()
|
|
|
|
return self.words[index]
|
|
|
|
|
|
|
|
|
|
|
|
def _apply_mask(docs, random_words, mask_prob=0.15):
|
|
|
|
# This needs to be here to avoid circular imports
|
|
|
|
from ...tokens import Doc
|
|
|
|
|
|
|
|
N = sum(len(doc) for doc in docs)
|
|
|
|
mask = numpy.random.uniform(0.0, 1.0, (N,))
|
|
|
|
mask = mask >= mask_prob
|
|
|
|
i = 0
|
|
|
|
masked_docs = []
|
|
|
|
for doc in docs:
|
|
|
|
words = []
|
|
|
|
for token in doc:
|
|
|
|
if not mask[i]:
|
|
|
|
word = _replace_word(token.text, random_words)
|
|
|
|
else:
|
|
|
|
word = token.text
|
|
|
|
words.append(word)
|
|
|
|
i += 1
|
|
|
|
spaces = [bool(w.whitespace_) for w in doc]
|
|
|
|
# NB: If you change this implementation to instead modify
|
|
|
|
# the docs in place, take care that the IDs reflect the original
|
|
|
|
# words. Currently we use the original docs to make the vectors
|
|
|
|
# for the target, so we don't lose the original tokens. But if
|
|
|
|
# you modified the docs in place here, you would.
|
|
|
|
masked_docs.append(Doc(doc.vocab, words=words, spaces=spaces))
|
|
|
|
return mask, masked_docs
|
|
|
|
|
|
|
|
|
|
|
|
def _replace_word(word, random_words, mask="[MASK]"):
|
|
|
|
roll = numpy.random.random()
|
|
|
|
if roll < 0.8:
|
|
|
|
return mask
|
|
|
|
elif roll < 0.9:
|
|
|
|
return random_words.next()
|
|
|
|
else:
|
2020-06-04 13:49:23 +00:00
|
|
|
return word
|