spaCy/spacy/pipeline/morphologizer.pyx
Vinit Ravishankar 28de85737f
Tagger label smoothing (#12293)
* add label smoothing

* use True/False instead of floats

* add entropy to debug data

* formatting

* docs

* change test to check difference in distributions

* Update website/docs/api/tagger.mdx

Co-authored-by: Adriane Boyd <adrianeboyd@gmail.com>

* Update spacy/pipeline/tagger.pyx

Co-authored-by: Adriane Boyd <adrianeboyd@gmail.com>

* bool -> float

* update docs

* fix seed

* black

* update tests to use label_smoothing = 0.0

* set default to 0.0, update quickstart

* Update spacy/pipeline/tagger.pyx

Co-authored-by: Adriane Boyd <adrianeboyd@gmail.com>

* update morphologizer, tagger test

* fix morph docs

* add url to docs

---------

Co-authored-by: Adriane Boyd <adrianeboyd@gmail.com>
2023-03-22 12:17:56 +01:00

312 lines
12 KiB
Cython

# cython: infer_types=True, profile=True, binding=True
from typing import Optional, Union, Dict, Callable
import srsly
from thinc.api import SequenceCategoricalCrossentropy, Model, Config
from itertools import islice
from ..tokens.doc cimport Doc
from ..vocab cimport Vocab
from ..morphology cimport Morphology
from ..parts_of_speech import IDS as POS_IDS
from ..symbols import POS
from ..language import Language
from ..errors import Errors
from .pipe import deserialize_config
from .tagger import Tagger
from .. import util
from ..scorer import Scorer
from ..training import validate_examples, validate_get_examples
from ..util import registry
# See #9050
BACKWARD_OVERWRITE = True
BACKWARD_EXTEND = False
default_model_config = """
[model]
@architectures = "spacy.Tagger.v2"
[model.tok2vec]
@architectures = "spacy.Tok2Vec.v2"
[model.tok2vec.embed]
@architectures = "spacy.CharacterEmbed.v2"
width = 128
rows = 7000
nM = 64
nC = 8
include_static_vectors = false
[model.tok2vec.encode]
@architectures = "spacy.MaxoutWindowEncoder.v2"
width = 128
depth = 4
window_size = 1
maxout_pieces = 3
"""
DEFAULT_MORPH_MODEL = Config().from_str(default_model_config)["model"]
@Language.factory(
"morphologizer",
assigns=["token.morph", "token.pos"],
default_config={"model": DEFAULT_MORPH_MODEL, "overwrite": True, "extend": False,
"scorer": {"@scorers": "spacy.morphologizer_scorer.v1"}, "label_smoothing": 0.0},
default_score_weights={"pos_acc": 0.5, "morph_acc": 0.5, "morph_per_feat": None},
)
def make_morphologizer(
nlp: Language,
model: Model,
name: str,
overwrite: bool,
extend: bool,
label_smoothing: float,
scorer: Optional[Callable],
):
return Morphologizer(nlp.vocab, model, name, overwrite=overwrite, extend=extend, label_smoothing=label_smoothing, scorer=scorer)
def morphologizer_score(examples, **kwargs):
def morph_key_getter(token, attr):
return getattr(token, attr).key
results = {}
results.update(Scorer.score_token_attr(examples, "pos", **kwargs))
results.update(Scorer.score_token_attr(examples, "morph", getter=morph_key_getter, **kwargs))
results.update(Scorer.score_token_attr_per_feat(examples,
"morph", getter=morph_key_getter, **kwargs))
return results
@registry.scorers("spacy.morphologizer_scorer.v1")
def make_morphologizer_scorer():
return morphologizer_score
class Morphologizer(Tagger):
POS_FEAT = "POS"
def __init__(
self,
vocab: Vocab,
model: Model,
name: str = "morphologizer",
*,
overwrite: bool = BACKWARD_OVERWRITE,
extend: bool = BACKWARD_EXTEND,
label_smoothing: float = 0.0,
scorer: Optional[Callable] = morphologizer_score,
):
"""Initialize a morphologizer.
vocab (Vocab): The shared vocabulary.
model (thinc.api.Model): The Thinc Model powering the pipeline component.
name (str): The component instance name, used to add entries to the
losses during training.
scorer (Optional[Callable]): The scoring method. Defaults to
Scorer.score_token_attr for the attributes "pos" and "morph" and
Scorer.score_token_attr_per_feat for the attribute "morph".
DOCS: https://spacy.io/api/morphologizer#init
"""
self.vocab = vocab
self.model = model
self.name = name
self._rehearsal_model = None
# to be able to set annotations without string operations on labels,
# store mappings from morph+POS labels to token-level annotations:
# 1) labels_morph stores a mapping from morph+POS->morph
# 2) labels_pos stores a mapping from morph+POS->POS
cfg = {
"labels_morph": {},
"labels_pos": {},
"overwrite": overwrite,
"extend": extend,
"label_smoothing": label_smoothing,
}
self.cfg = dict(sorted(cfg.items()))
self.scorer = scorer
@property
def labels(self):
"""RETURNS (Tuple[str]): The labels currently added to the component."""
return tuple(self.cfg["labels_morph"].keys())
@property
def label_data(self) -> Dict[str, Dict[str, Union[str, float, int, None]]]:
"""A dictionary with all labels data."""
return {"morph": self.cfg["labels_morph"], "pos": self.cfg["labels_pos"]}
def add_label(self, label):
"""Add a new label to the pipe.
label (str): The label to add.
RETURNS (int): 0 if label is already present, otherwise 1.
DOCS: https://spacy.io/api/morphologizer#add_label
"""
if not isinstance(label, str):
raise ValueError(Errors.E187)
if label in self.labels:
return 0
self._allow_extra_label()
# normalize label
norm_label = self.vocab.morphology.normalize_features(label)
# extract separate POS and morph tags
label_dict = Morphology.feats_to_dict(label)
pos = label_dict.get(self.POS_FEAT, "")
if self.POS_FEAT in label_dict:
label_dict.pop(self.POS_FEAT)
# normalize morph string and add to morphology table
norm_morph = self.vocab.strings[self.vocab.morphology.add(label_dict)]
# add label mappings
if norm_label not in self.cfg["labels_morph"]:
self.cfg["labels_morph"][norm_label] = norm_morph
self.cfg["labels_pos"][norm_label] = POS_IDS[pos]
return 1
def initialize(self, get_examples, *, nlp=None, labels=None):
"""Initialize the pipe for training, using a representative set
of data examples.
get_examples (Callable[[], Iterable[Example]]): Function that
returns a representative sample of gold-standard Example objects.
nlp (Language): The current nlp object the component is part of.
DOCS: https://spacy.io/api/morphologizer#initialize
"""
validate_get_examples(get_examples, "Morphologizer.initialize")
util.check_lexeme_norms(self.vocab, "morphologizer")
if labels is not None:
self.cfg["labels_morph"] = labels["morph"]
self.cfg["labels_pos"] = labels["pos"]
else:
# First, fetch all labels from the data
for example in get_examples():
for i, token in enumerate(example.reference):
pos = token.pos_
# if both are unset, annotation is missing, so do not add
# an empty label
if pos == "" and not token.has_morph():
continue
morph = str(token.morph)
# create and add the combined morph+POS label
morph_dict = Morphology.feats_to_dict(morph)
if pos:
morph_dict[self.POS_FEAT] = pos
norm_label = self.vocab.strings[self.vocab.morphology.add(morph_dict)]
# add label->morph and label->POS mappings
if norm_label not in self.cfg["labels_morph"]:
self.cfg["labels_morph"][norm_label] = morph
self.cfg["labels_pos"][norm_label] = POS_IDS[pos]
if len(self.labels) < 1:
raise ValueError(Errors.E143.format(name=self.name))
doc_sample = []
label_sample = []
for example in islice(get_examples(), 10):
gold_array = []
for i, token in enumerate(example.reference):
pos = token.pos_
morph = str(token.morph)
morph_dict = Morphology.feats_to_dict(morph)
if pos:
morph_dict[self.POS_FEAT] = pos
norm_label = self.vocab.strings[self.vocab.morphology.add(morph_dict)]
gold_array.append([1.0 if label == norm_label else 0.0 for label in self.labels])
doc_sample.append(example.x)
label_sample.append(self.model.ops.asarray(gold_array, dtype="float32"))
assert len(doc_sample) > 0, Errors.E923.format(name=self.name)
assert len(label_sample) > 0, Errors.E923.format(name=self.name)
self.model.initialize(X=doc_sample, Y=label_sample)
def set_annotations(self, docs, batch_tag_ids):
"""Modify a batch of documents, using pre-computed scores.
docs (Iterable[Doc]): The documents to modify.
batch_tag_ids: The IDs to set, produced by Morphologizer.predict.
DOCS: https://spacy.io/api/morphologizer#set_annotations
"""
if isinstance(docs, Doc):
docs = [docs]
cdef Doc doc
cdef Vocab vocab = self.vocab
cdef bint overwrite = self.cfg["overwrite"]
cdef bint extend = self.cfg["extend"]
labels = self.labels
for i, doc in enumerate(docs):
doc_tag_ids = batch_tag_ids[i]
if hasattr(doc_tag_ids, "get"):
doc_tag_ids = doc_tag_ids.get()
for j, tag_id in enumerate(doc_tag_ids):
morph = labels[tag_id]
# set morph
if doc.c[j].morph == 0 or overwrite or extend:
if overwrite and extend:
# morphologizer morph overwrites any existing features
# while extending
extended_morph = Morphology.feats_to_dict(self.vocab.strings[doc.c[j].morph])
extended_morph.update(Morphology.feats_to_dict(self.cfg["labels_morph"].get(morph, 0)))
doc.c[j].morph = self.vocab.morphology.add(extended_morph)
elif extend:
# existing features are preserved and any new features
# are added
extended_morph = Morphology.feats_to_dict(self.cfg["labels_morph"].get(morph, 0))
extended_morph.update(Morphology.feats_to_dict(self.vocab.strings[doc.c[j].morph]))
doc.c[j].morph = self.vocab.morphology.add(extended_morph)
else:
# clobber
doc.c[j].morph = self.vocab.morphology.add(self.cfg["labels_morph"].get(morph, 0))
# set POS
if doc.c[j].pos == 0 or overwrite:
doc.c[j].pos = self.cfg["labels_pos"].get(morph, 0)
def get_loss(self, examples, scores):
"""Find the loss and gradient of loss for the batch of documents and
their predicted scores.
examples (Iterable[Examples]): The batch of examples.
scores: Scores representing the model's predictions.
RETURNS (Tuple[float, float]): The loss and the gradient.
DOCS: https://spacy.io/api/morphologizer#get_loss
"""
validate_examples(examples, "Morphologizer.get_loss")
loss_func = SequenceCategoricalCrossentropy(names=self.labels, normalize=False,
label_smoothing=self.cfg["label_smoothing"])
truths = []
for eg in examples:
eg_truths = []
pos_tags = eg.get_aligned("POS", as_string=True)
morphs = eg.get_aligned("MORPH", as_string=True)
for i in range(len(morphs)):
pos = pos_tags[i]
morph = morphs[i]
# POS may align (same value for multiple tokens) when morph
# doesn't, so if either is misaligned (None), treat the
# annotation as missing so that truths doesn't end up with an
# unknown morph+POS combination
if pos is None or morph is None:
label = None
# If both are unset, the annotation is missing (empty morph
# converted from int is "_" rather than "")
elif pos == "" and morph == "":
label = None
# Otherwise, generate the combined label
else:
label_dict = Morphology.feats_to_dict(morph)
if pos:
label_dict[self.POS_FEAT] = pos
label = self.vocab.strings[self.vocab.morphology.add(label_dict)]
# As a fail-safe, skip any unrecognized labels
if label not in self.labels:
label = None
eg_truths.append(label)
truths.append(eg_truths)
d_scores, loss = loss_func(scores, truths)
if self.model.ops.xp.isnan(loss):
raise ValueError(Errors.E910.format(name=self.name))
return float(loss), d_scores