10 KiB
title | teaser | menu | ||||||
---|---|---|---|---|---|---|---|---|
What's New in v2.1 | New features, backwards incompatibilities and migration guide |
|
New Features
spaCy v2.1 has focussed primarily on stability and performance, solidifying the design changes introduced in v2.0. As well as smaller models, faster runtime, and many bug-fixes, v2.1 also introduces experimental support for some exciting new NLP innovations. For the full changelog, see the release notes on GitHub.
BERT/ULMFit/Elmo-style pre-training
Example
$ python -m spacy pretrain ./raw_text.jsonl en_vectors_web_lg ./pretrained-model
spaCy v2.1 introduces a new CLI command, spacy pretrain
, that can make your
models much more accurate. It's especially useful when you have limited
training data. The spacy pretrain
command lets you use transfer learning to
initialize your models with information from raw text, using a language model
objective similar to the one used in Google's BERT system. We've taken
particular care to ensure that pretraining works well even with spaCy's small
default architecture sizes, so you don't have to compromise on efficiency to use
it.
API: spacy pretrain
**Usage: **
Improving accuracy with transfer learning
Extended match pattern API
Example
# Matches "love cats" or "likes flowers" pattern1 = [{"LEMMA": {"IN": ["like", "love"]}}, {"POS": "NOUN"}] # Matches tokens of length >= 10 pattern2 = [{"LENGTH": {">=": 10}}] # Matches custom attribute with regex pattern3 = [{"_": {"country": {"REGEX": "^([Uu](\\.?|nited) ?[Ss](\\.?|tates)"}}}]
Instead of mapping to a single value, token patterns can now also map to a
dictionary of properties. For example, to specify that the value of a lemma
should be part of a list of values, or to set a minimum character length. It now
also supports a REGEX
property, as well as set membership via IN
and
NOT_IN
, custom extension attributes via _
and rich comparison for numeric
values.
API: Matcher
**Usage: **
Extended pattern syntax and attributes,
Regular expressions
Easy rule-based entity recognition
Example
from spacy.pipeline import EntityRuler ruler = EntityRuler(nlp) ruler.add_patterns([{"label": "ORG", "pattern": "Apple"}]) nlp.add_pipe(ruler, before="ner")
The EntityRuler
is an exciting new component that lets you add named entities
based on pattern dictionaries, and makes it easy to combine rule-based and
statistical named entity recognition for even more powerful models. Entity rules
can be phrase patterns for exact string matches, or token patterns for full
flexibility.
API: EntityRuler
**Usage: **
Rule-based entity recognition
Phrase matching with other attributes
Example
matcher = PhraseMatcher(nlp.vocab, attr="POS") matcher.add("PATTERN", None, nlp(u"I love cats")) doc = nlp(u"You like dogs") matches = matcher(doc)
By default, the PhraseMatcher
will match on the verbatim token text, e.g.
Token.text
. By setting the attr
argument on initialization, you can change
which token attribute the matcher should use when comparing the phrase
pattern to the matched Doc
. For example, LOWER
for case-insensitive matches
or POS
for finding sequences of the same part-of-speech tags.
API: PhraseMatcher
**Usage: **
Matching on other token attributes
Components and languages via entry points
Example
from setuptools import setup setup( name="custom_extension_package", entry_points={ "spacy_factories": ["your_component = component:ComponentFactory"] "spacy_languages": ["xyz = language:XYZLanguage"] } )
Using entry points, model packages and extension packages can now define their
own "spacy_factories"
and "spacy_languages"
, which will be added to the
built-in factories and languages. If a package in the same environment exposes
spaCy entry points, all of this happens automatically and no further user action
is required.
Usage: Using entry points
Retokenizer for merging and splitting
Example
doc = nlp(u"I like David Bowie") with doc.retokenize() as retokenizer: attrs = {"LEMMA": u"David Bowie"} retokenizer.merge(doc[2:4], attrs=attrs)
The new Doc.retokenize
context manager allows merging spans of multiple tokens
into one single token, and splitting single tokens into multiple tokens.
Modifications to the Doc
's tokenization are stored, and then made all at once
when the context manager exits. This is much more efficient, and less
error-prone. Doc.merge
and Span.merge
still work, but they're considered
deprecated.
API: Doc.retokenize
,
Retokenizer.merge
,
Retokenizer.split
**Usage:
**Merging and splitting
Improved documentation
Although it looks pretty much the same, we've rebuilt the entire documentation using Gatsby and MDX. It's now an even faster progressive web app and allows us to write all content entirely in Markdown, without having to compromise on easy-to-use custom UI components. We're hoping that the Markdown source will make it even easier to contribute to the documentation. For more details, check out the styleguide and source. While converting the pages to Markdown, we've also fixed a bunch of typos, improved the existing pages and added some new content:
- Usage Guide: Rule-based Matching
How to use theMatcher
,PhraseMatcher
and the newEntityRuler
, and write powerful components to combine statistical models and rules. - Usage Guide: Saving and Loading
Everything you need to know about serialization, and how to save and load pipeline components, package your spaCy models as Python modules and use entry points. - **Usage Guide: **
Merging and Splitting
How to retokenize aDoc
using the newretokenize
context manager and merge spans into single tokens and split single tokens into multiple. - Universe: Videos and Podcasts
- API:
EntityRuler
- API:
SentenceSegmenter
- API: Pipeline functions
Backwards incompatibilities
If you've been training your own models, you'll need to retrain them
with the new version. Also don't forget to upgrade all models to the latest
versions. Models for v2.0.x aren't compatible with models for v2.1.x. To check
if all of your models are up to date, you can run the
spacy validate
command.
-
While the
Matcher
API is fully backwards compatible, its algorithm has changed to fix a number of bugs and performance issues. This means that theMatcher
in v2.1.x may produce different results compared to theMatcher
in v2.0.x. -
For better compatibility with the Universal Dependencies data, the lemmatizer now preserves capitalization, e.g. for proper nouns. See this issue for details.
-
The built-in rule-based sentence boundary detector is now only called
"sentencizer"
– the name"sbd"
is deprecated.- sentence_splitter = nlp.create_pipe("sbd") + sentence_splitter = nlp.create_pipe("sentencizer")
-
The keyword argument
n_threads
on the.pipe
methods is now deprecated, as the v2.x models cannot release the global interpreter lock. (Future versions may introduce an_process
argument for parallel inference via multiprocessing.) -
The
Doc.print_tree
method is now deprecated. If you need a custom nested JSON representation of aDoc
object, you might want to write your own helper function. For a simple and consistent JSON representation of theDoc
object and its annotations, you can now use theDoc.to_json
method. Going forward, this method will output the same format as the JSON training data expected byspacy train
. -
The
spacy train
command now lets you specify a comma-separated list of pipeline component names, instead of separate flags like--no-parser
to disable components. This is more flexible and also handles custom components out-of-the-box.- $ spacy train en /output train_data.json dev_data.json --no-parser + $ spacy train en /output train_data.json dev_data.json --pipeline tagger,ner
-
The
spacy init-model
command now uses a--jsonl-loc
argument to pass in a a newline-delimited JSON (JSONL) file containing one lexical entry per line instead of a separate--freqs-loc
and--clusters-loc
.- $ spacy init-model en ./model --freqs-loc ./freqs.txt --clusters-loc ./clusters.txt + $ spacy init-model en ./model --jsonl-loc ./vocab.jsonl
-
Also note that some of the model licenses have changed:
it_core_news_sm
is now correctly licensed under CC BY-NC-SA 3.0, and all English and German models are now published under the MIT license.