def test_language_wordpiece_to_from_bytes(name):
nlp = PyTT_Language()
nlp.add_pipe(nlp.create_pipe("sentencizer"))
wordpiecer = PyTT_WordPiecer.from_pretrained(nlp.vocab, pytt_name=name)
nlp.add_pipe(wordpiecer)
doc = nlp("hello world")
assert doc._.pytt_word_pieces is not None
nlp2 = PyTT_Language()
nlp2.add_pipe(nlp.create_pipe("sentencizer"))
nlp2.add_pipe(PyTT_WordPiecer(nlp2.vocab))
with pytest.raises(ValueError):
new_doc = nlp2("hello world")
nlp2.from_bytes(nlp.to_bytes())
new_doc = nlp2("hello world")
assert new_doc._.pytt_word_pieces is not None
def nlp(name):
p_nlp = PyTT_Language(pytt_name=name)
p_nlp.add_pipe(p_nlp.create_pipe("sentencizer"))
p_nlp.add_pipe(PyTT_WordPiecer.from_pretrained(p_nlp.vocab,
pytt_name=name))
p_nlp.add_pipe(
PyTT_TokenVectorEncoder.from_pretrained(p_nlp.vocab, name=name))
return p_nlp
class SpacyTokenToVec(TokenToVec):
def __init__(self, model_name: str, model_path: Optional[str]) -> None:
super().__init__()
with SwitchDefaultTensor():
if model_path:
self.nlp = PyTT_Language(pytt_name=model_name,
meta={"lang": "en"})
self.nlp.add_pipe(self.nlp.create_pipe("sentencizer"))
self.nlp.add_pipe(
PyTT_WordPiecer.from_pretrained(self.nlp.vocab,
model_name))
self.nlp.add_pipe(
PyTT_TokenVectorEncoder.from_pretrained(
self.nlp.vocab, model_path))
else:
self.nlp = spacy.load(model_name)
if not model_name in NAME_TO_DIM:
raise ValueError("Model name is unknown, I know " +
str(list(NAME_TO_DIM.keys())))
self.output_dim = NAME_TO_DIM[model_name]
def get_output_dim(self) -> int:
return self.output_dim
def embed(self, vocab: Vocabulary, tokens: torch.Tensor) -> torch.Tensor:
"""
Idea: reconstruct string tokens from token ids -> feed to spacy -> return tensors
:param vocab:
:param tokens:
:return:
"""
with SwitchDefaultTensor():
embedded_sentences = []
tokens_cpu = tokens.cpu()
batch_size, seq_len = tokens.shape
for sentence in tokens_cpu:
str_tokens: List[str] = [
vocab.get_token_from_index(int(token))
for token in sentence if token != 0
] #skip padding
doc = Doc(self.nlp.vocab, words=str_tokens)
self.nlp.pipeline[1][1](doc) #word pieces
self.nlp.pipeline[2][1](doc) #run transformer on wordpieces
#add padding back in
#embedded = torch.from_numpy(cupy.asnumpy(doc.tensor)).to(device) # shape (str_tokens, output dim)
embedded = from_dlpack(
doc.tensor.toDlpack()) # shape (str_tokens, output dim)
assert embedded.shape == (len(str_tokens),
self.get_output_dim())
if seq_len - len(str_tokens) > 0:
padded = torch.zeros(seq_len - len(str_tokens),
self.get_output_dim())
embedded = torch.cat([embedded, padded], dim=0)
embedded_sentences.append(embedded)
return torch.stack(embedded_sentences, dim=0)
def test_language_wordpiece_tok2vec_to_from_bytes(nlp, name):
doc = nlp("hello world")
assert is_valid_tensor(doc.tensor)
nlp2 = PyTT_Language()
nlp2.add_pipe(nlp2.create_pipe("sentencizer"))
nlp2.add_pipe(PyTT_WordPiecer(nlp.vocab))
nlp2.add_pipe(PyTT_TokenVectorEncoder(nlp.vocab))
with pytest.raises(ValueError):
new_doc = nlp2("hello world")
nlp2.from_bytes(nlp.to_bytes())
new_doc = nlp2("hello world")
assert is_valid_tensor(new_doc.tensor)
assert new_doc._.pytt_word_pieces is not None
def main(path, name="bert-base-uncased", lang="en"):
msg = Printer()
msg.info(f"Creating model for '{name}' ({lang})")
with msg.loading(f"Setting up the pipeline..."):
nlp = PyTT_Language(pytt_name=name, meta={"lang": lang})
nlp.add_pipe(nlp.create_pipe("sentencizer"))
nlp.add_pipe(PyTT_WordPiecer.from_pretrained(nlp.vocab, name))
nlp.add_pipe(PyTT_TokenVectorEncoder.from_pretrained(nlp.vocab, name))
msg.good("Initialized the model pipeline")
nlp.to_disk(path)
msg.good(f"Saved '{name}' ({lang})")
msg.text(f"Pipeline: {nlp.pipe_names}")
msg.text(f"Location: {path}")
def test_language_to_from_disk(nlp, name):
doc = nlp("hello world")
assert is_valid_tensor(doc.tensor)
with make_tempdir() as tempdir:
nlp.to_disk(tempdir)
new_nlp = PyTT_Language()
new_nlp.add_pipe(new_nlp.create_pipe("sentencizer"))
wordpiecer = PyTT_WordPiecer(new_nlp.vocab, pytt_name=name)
tok2vec = PyTT_TokenVectorEncoder(new_nlp.vocab, pytt_name=name)
new_nlp.add_pipe(wordpiecer)
new_nlp.add_pipe(tok2vec)
new_nlp.from_disk(tempdir)
assert new_nlp.pipe_names == nlp.pipe_names
new_doc = new_nlp("hello world")
assert is_valid_tensor(new_doc.tensor)
assert_equal(doc.tensor, new_doc.tensor)
from spacy_pytorch_transformers import PyTT_Language, PyTT_WordPiecer, PyTT_TokenVectorEncoder
from pathlib import Path
pytorch_path = str(Path.home() / "pytorch-rubert")
spacy_path = str(Path.home() / "spacy-rubert")
name = "ru_pytt_rubert_cased"
nlp = PyTT_Language(pytt_name=name, meta={"lang": "ru"})
nlp.add_pipe(nlp.create_pipe("sentencizer"))
nlp.add_pipe(PyTT_WordPiecer.from_pretrained(nlp.vocab, pytorch_path))
nlp.add_pipe(PyTT_TokenVectorEncoder.from_pretrained(nlp.vocab, pytorch_path))
print(nlp.pipe_names)
nlp.to_disk(spacy_path)
class BatchedSpacyTokenToVec(TokenToVec):
def __init__(self,
model_name: str,
model_path: Optional[str] = None) -> None:
"""
Loads a model_name (e.g. en_pytt_xlnetbasecased_lg) or a combination
of model name and local model path (e.g. xlnet-large-cased and /local/mlinde/xlnet-large-cased)
see https://github.com/explosion/spacy-pytorch-transformers#loading-models-from-a-path for how to prepare a model
:param model_name:
:param model_path:
"""
super().__init__()
with SwitchDefaultTensor():
if model_path:
self.nlp = PyTT_Language(pytt_name=model_name,
meta={"lang": "en"})
self.nlp.add_pipe(self.nlp.create_pipe("sentencizer"))
self.nlp.add_pipe(
PyTT_WordPiecer.from_pretrained(self.nlp.vocab,
model_name))
self.nlp.add_pipe(
PyTT_TokenVectorEncoder.from_pretrained(
self.nlp.vocab, model_path))
else:
self.nlp = spacy.load(model_name)
if model_name not in NAME_TO_DIM:
raise ValueError("Model name is unknown, I know " +
str(list(NAME_TO_DIM.keys())))
self.output_dim = NAME_TO_DIM[model_name]
def get_output_dim(self) -> int:
return self.output_dim
def embed(self, vocab: Vocabulary, tokens: torch.Tensor) -> torch.Tensor:
"""
Idea: reconstruct string tokens from token ids -> feed to spacy -> return tensors
:param vocab:
:param tokens:
:return:
"""
with SwitchDefaultTensor():
with torch.autograd.no_grad():
embedded_sentences = []
tokens_cpu = tokens.cpu()
batch_size, seq_len = tokens.shape
sents = []
for sentence in tokens_cpu:
str_tokens: List[str] = [
vocab.get_token_from_index(int(token))
for token in sentence if token != 0
] #skip padding
sents.append(str_tokens)
doc = make_doc(self.nlp.vocab, sents)
self.nlp.pipeline[1][1](doc) #word pieces
self.nlp.pipeline[2][1](doc) #run transformer on wordpieces
#Now iterate over sentences in correct order and cut out the correct tensor + pad it
for sent, str_tokens in zip(doc.sents, sents):
#add padding back in
embedded = from_dlpack(sent.tensor.toDlpack()
) # shape (str_tokens, output dim)
if seq_len - len(str_tokens) > 0:
padded = torch.zeros(seq_len - len(str_tokens),
self.get_output_dim())
embedded = torch.cat([embedded, padded], dim=0)
embedded_sentences.append(embedded)
return torch.stack(embedded_sentences, dim=0)