def build_vocab(self, *args, **kwargs):
"""Add unaligned_token to the list of special symbols."""
counter = Counter()
sources = []
for arg in args:
if isinstance(arg, data.Dataset):
sources += [
getattr(arg, name) for name, field in arg.fields.items()
if field is self
]
else:
sources.append(arg)
for sample in sources:
for x in sample:
if not self.sequential:
x = [x]
try:
counter.update(x)
except TypeError:
counter.update(chain.from_iterable(x))
specials = list(
OrderedDict.fromkeys(tok for tok in [
self.unk_token,
self.pad_token,
self.init_token,
self.eos_token,
self.unaligned_token,
] if tok is not None))
self.vocab = self.vocab_cls(counter, specials=specials, **kwargs)
def build_vocab(self, *args, **kwargs):
"""Add unaligned_token to the list of special symbols."""
counter = Counter()
sources = []
for arg in args: # arg是QEDataset类，里面包括examples和fields
if isinstance(arg, data.Dataset):
sources += [
getattr(arg, name) for name, field in arg.fields.items()
if field is self
] # source是列表，列表中元素是迭代器
else:
sources.append(arg)
for sample in sources:
for x in sample: # 每次循环读取一个样本，将样本处理成list形式，然后更新counter
if not self.sequential:
x = [x]
try:
counter.update(x)
except TypeError:
counter.update(chain.from_iterable(x))
specials = list(
OrderedDict.fromkeys(tok for tok in [
self.unk_token, self.pad_token, self.init_token,
self.eos_token, self.unaligned_token
] if tok is not None)) # ['<unk>', '<pad>', '<unaligned>']
self.vocab = self.vocab_cls(counter, specials=specials, **kwargs)
def fit_vocab(
self,
samples,
vocab_size=None,
vocab_min_freq=0,
embeddings_name=None,
keep_rare_words_with_embeddings=False,
add_embeddings_vocab=False,
):
tokens = Counter()
for sample in samples:
# TODO: subtokenize?
tokens.update(self.tokenize(sample))
# We use our own Vocabulary class
specials = list(
OrderedDict.fromkeys(
tok for tok in [self.unaligned_token] if tok is not None
)
)
# TODO: handle embeddings/vectors
self.vocab = Vocabulary(
tokens,
max_size=vocab_size,
min_freq=vocab_min_freq,
unk_token=self.unk_token,
pad_token=self.pad_token,
bos_token=self.bos_token,
eos_token=self.eos_token,
specials=specials,
specials_first=self.specials_first,
# TODO: missing vectors, etc.
vectors=None,
rare_with_vectors=keep_rare_words_with_embeddings,
add_vectors_vocab=add_embeddings_vocab,
)
