mirror of
https://github.com/explosion/spaCy.git
synced 2024-11-13 21:26:58 +03:00
251 lines
11 KiB
Plaintext
251 lines
11 KiB
Plaintext
//- 💫 DOCS > USAGE > WHAT'S NEW IN V2.0 > NEW FEATURES
|
||
|
||
p
|
||
| This section contains an overview of the most important
|
||
| #[strong new features and improvements]. The #[+a("/api") API docs]
|
||
| include additional deprecation notes. New methods and functions that
|
||
| were introduced in this version are marked with a
|
||
| #[span.u-text-tag.u-text-tag--spaced v2.0] tag.
|
||
|
||
+h(3, "features-models") Convolutional neural network models
|
||
|
||
+aside-code("Example", "bash")
|
||
for _, lang in MODELS
|
||
if lang != "xx"
|
||
| python -m spacy download #{lang} # default #{LANGUAGES[lang]} model!{'\n'}
|
||
| python -m spacy download xx_ent_wiki_sm # multi-language NER
|
||
|
||
p
|
||
| spaCy v2.0 features new neural models for tagging,
|
||
| parsing and entity recognition. The models have
|
||
| been designed and implemented from scratch specifically for spaCy, to
|
||
| give you an unmatched balance of speed, size and accuracy. The new
|
||
| models are #[strong 10× smaller], #[strong 20% more accurate],
|
||
| and #[strong even cheaper to run] than the previous generation.
|
||
|
||
p
|
||
| spaCy v2.0's new neural network models bring significant improvements in
|
||
| accuracy, especially for English Named Entity Recognition. The new
|
||
| #[+a("/models/en#en_core_web_lg") #[code en_core_web_lg]] model makes
|
||
| about #[strong 25% fewer mistakes] than the corresponding v1.x model and
|
||
| is within #[strong 1% of the current state-of-the-art]
|
||
| (#[+a("https://arxiv.org/pdf/1702.02098.pdf") Strubell et al., 2017]).
|
||
| The v2.0 models are also cheaper to run at scale, as they require
|
||
| #[strong under 1 GB of memory] per process.
|
||
|
||
+infobox
|
||
| #[+label-inline Usage:] #[+a("/models") Models directory],
|
||
| #[+a("/models/comparison") Models comparison],
|
||
| #[+a("#benchmarks") Benchmarks]
|
||
|
||
+h(3, "features-pipelines") Improved processing pipelines
|
||
|
||
+aside-code("Example").
|
||
# Set custom attributes
|
||
Doc.set_extension('my_attr', default=False)
|
||
Token.set_extension('my_attr', getter=my_token_getter)
|
||
assert doc._.my_attr, token._.my_attr
|
||
|
||
# Add components to the pipeline
|
||
my_component = lambda doc: doc
|
||
nlp.add_pipe(my_component)
|
||
|
||
p
|
||
| It's now much easier to #[strong customise the pipeline] with your own
|
||
| components: functions that receive a #[code Doc] object, modify and
|
||
| return it. Extensions let you write any
|
||
| #[strong attributes, properties and methods] to the #[code Doc],
|
||
| #[code Token] and #[code Span]. You can add data, implement new
|
||
| features, integrate other libraries with spaCy or plug in your own
|
||
| machine learning models.
|
||
|
||
+image
|
||
include ../../assets/img/pipeline.svg
|
||
|
||
+infobox
|
||
| #[+label-inline API:] #[+api("language") #[code Language]],
|
||
| #[+api("doc#set_extension") #[code Doc.set_extension]],
|
||
| #[+api("span#set_extension") #[code Span.set_extension]],
|
||
| #[+api("token#set_extension") #[code Token.set_extension]]
|
||
| #[+label-inline Usage:]
|
||
| #[+a("/usage/processing-pipelines") Processing pipelines]
|
||
| #[+label-inline Code:]
|
||
| #[+src("/usage/examples#section-pipeline") Pipeline examples]
|
||
|
||
+h(3, "features-text-classification") Text classification
|
||
|
||
+aside-code("Example").
|
||
textcat = nlp.create_pipe('textcat')
|
||
nlp.add_pipe(textcat, last=True)
|
||
optimizer = nlp.begin_training()
|
||
for itn in range(100):
|
||
for doc, gold in train_data:
|
||
nlp.update([doc], [gold], sgd=optimizer)
|
||
doc = nlp(u'This is a text.')
|
||
print(doc.cats)
|
||
|
||
p
|
||
| spaCy v2.0 lets you add text categorization models to spaCy pipelines.
|
||
| The model supports classification with multiple, non-mutually
|
||
| exclusive labels – so multiple labels can apply at once. You can
|
||
| change the model architecture rather easily, but by default, the
|
||
| #[code TextCategorizer] class uses a convolutional neural network to
|
||
| assign position-sensitive vectors to each word in the document.
|
||
|
||
+infobox
|
||
| #[+label-inline API:] #[+api("textcategorizer") #[code TextCategorizer]],
|
||
| #[+api("doc#attributes") #[code Doc.cats]],
|
||
| #[+api("goldparse#attributes") #[code GoldParse.cats]]#[br]
|
||
| #[+label-inline Usage:] #[+a("/usage/training#textcat") Training a text classication model]
|
||
|
||
+h(3, "features-hash-ids") Hash values instead of integer IDs
|
||
|
||
+aside-code("Example").
|
||
doc = nlp(u'I love coffee')
|
||
assert doc.vocab.strings[u'coffee'] == 3197928453018144401
|
||
assert doc.vocab.strings[3197928453018144401] == u'coffee'
|
||
|
||
beer_hash = doc.vocab.strings.add(u'beer')
|
||
assert doc.vocab.strings[u'beer'] == beer_hash
|
||
assert doc.vocab.strings[beer_hash] == u'beer'
|
||
|
||
p
|
||
| The #[+api("stringstore") #[code StringStore]] now resolves all strings
|
||
| to hash values instead of integer IDs. This means that the string-to-int
|
||
| mapping #[strong no longer depends on the vocabulary state], making a lot
|
||
| of workflows much simpler, especially during training. Unlike integer IDs
|
||
| in spaCy v1.x, hash values will #[strong always match] – even across
|
||
| models. Strings can now be added explicitly using the new
|
||
| #[+api("stringstore#add") #[code Stringstore.add]] method. A token's hash
|
||
| is available via #[code token.orth].
|
||
|
||
+infobox
|
||
| #[+label-inline API:] #[+api("stringstore") #[code StringStore]]
|
||
| #[+label-inline Usage:] #[+a("/usage/spacy-101#vocab") Vocab, hashes and lexemes 101]
|
||
|
||
+h(3, "features-vectors") Improved word vectors support
|
||
|
||
+aside-code("Example").
|
||
for word, vector in vector_data:
|
||
nlp.vocab.set_vector(word, vector)
|
||
nlp.vocab.vectors.from_glove('/path/to/vectors')
|
||
# keep 10000 unique vectors and remap the rest
|
||
nlp.vocab.prune_vectors(10000)
|
||
nlp.to_disk('/model')
|
||
|
||
p
|
||
| The new #[+api("vectors") #[code Vectors]] class helps the
|
||
| #[code Vocab] manage the vectors assigned to strings, and lets you
|
||
| assign vectors individually, or
|
||
| #[+a("/usage/vectors-similarity#custom-loading-glove") load in GloVe vectors]
|
||
| from a directory. To help you strike a good balance between coverage
|
||
| and memory usage, the #[code Vectors] class lets you map
|
||
| #[strong multiple keys] to the #[strong same row] of the table. If
|
||
| you're using the #[+api("cli#vocab") #[code spacy vocab]] command to
|
||
| create a vocabulary, pruning the vectors will be taken care of
|
||
| automatically. Otherwise, you can use the new
|
||
| #[+api("vocab#prune_vectors") #[code Vocab.prune_vectors]].
|
||
|
||
+infobox
|
||
| #[+label-inline API:] #[+api("vectors") #[code Vectors]],
|
||
| #[+api("vocab") #[code Vocab]]
|
||
| #[+label-inline Usage:] #[+a("/usage/vectors-similarity") Word vectors and semantic similarity]
|
||
|
||
+h(3, "features-serializer") Saving, loading and serialization
|
||
|
||
+aside-code("Example").
|
||
nlp = spacy.load('en') # shortcut link
|
||
nlp = spacy.load('en_core_web_sm') # package
|
||
nlp = spacy.load('/path/to/en') # unicode path
|
||
nlp = spacy.load(Path('/path/to/en')) # pathlib Path
|
||
|
||
nlp.to_disk('/path/to/nlp')
|
||
nlp = English().from_disk('/path/to/nlp')
|
||
|
||
p
|
||
| spaCy's serialization API has been made consistent across classes and
|
||
| objects. All container classes, i.e. #[code Language], #[code Doc],
|
||
| #[code Vocab] and #[code StringStore] now have a #[code to_bytes()],
|
||
| #[code from_bytes()], #[code to_disk()] and #[code from_disk()] method
|
||
| that supports the Pickle protocol.
|
||
|
||
p
|
||
| The improved #[code spacy.load] makes loading models easier and more
|
||
| transparent. You can load a model by supplying its
|
||
| #[+a("/usage/models#usage") shortcut link], the name of an installed
|
||
| #[+a("/models") model package] or a path. The #[code Language] class to
|
||
| initialise will be determined based on the model's settings. For a blank l
|
||
| anguage, you can import the class directly, e.g.
|
||
| #[code.u-break from spacy.lang.en import English] or use
|
||
| #[+api("spacy#blank") #[code spacy.blank()]].
|
||
|
||
+infobox
|
||
| #[+label-inline API:] #[+api("spacy#load") #[code spacy.load]],
|
||
| #[+api("language#to_disk") #[code Language.to_disk]]
|
||
| #[+label-inline Usage:] #[+a("/usage/models#usage") Models],
|
||
| #[+a("/usage/training#saving-loading") Saving and loading]
|
||
|
||
+h(3, "features-displacy") displaCy visualizer with Jupyter support
|
||
|
||
+aside-code("Example").
|
||
from spacy import displacy
|
||
doc = nlp(u'This is a sentence about Facebook.')
|
||
displacy.serve(doc, style='dep') # run the web server
|
||
html = displacy.render(doc, style='ent') # generate HTML
|
||
|
||
p
|
||
| Our popular dependency and named entity visualizers are now an official
|
||
| part of the spaCy library. displaCy can run a simple web server, or
|
||
| generate raw HTML markup or SVG files to be exported. You can pass in one
|
||
| or more docs, and customise the style. displaCy also auto-detects whether
|
||
| you're running #[+a("https://jupyter.org") Jupyter] and will render the
|
||
| visualizations in your notebook.
|
||
|
||
+infobox
|
||
| #[+label-inline API:] #[+api("top-level#displacy") #[code displacy]]
|
||
| #[+label-inline Usage:] #[+a("/usage/visualizers") Visualizing spaCy]
|
||
|
||
+h(3, "features-language") Improved language data and lazy loading
|
||
|
||
p
|
||
| Language-specfic data now lives in its own submodule, #[code spacy.lang].
|
||
| Languages are lazy-loaded, i.e. only loaded when you import a
|
||
| #[code Language] class, or load a model that initialises one. This allows
|
||
| languages to contain more custom data, e.g. lemmatizer lookup tables, or
|
||
| complex regular expressions. The language data has also been tidied up
|
||
| and simplified. spaCy now also supports simple lookup-based
|
||
| lemmatization – and #[strong #{LANG_COUNT} languages] in total!
|
||
|
||
+infobox
|
||
| #[+label-inline API:] #[+api("language") #[code Language]]
|
||
| #[+label-inline Code:] #[+src(gh("spaCy", "spacy/lang")) #[code spacy/lang]]
|
||
| #[+label-inline Usage:] #[+a("/usage/adding-languages") Adding languages]
|
||
|
||
+h(3, "features-matcher") Revised matcher API and phrase matcher
|
||
|
||
+aside-code("Example").
|
||
from spacy.matcher import Matcher, PhraseMatcher
|
||
|
||
matcher = Matcher(nlp.vocab)
|
||
matcher.add('HEARTS', None, [{'ORTH': '❤️', 'OP': '+'}])
|
||
|
||
phrasematcher = PhraseMatcher(nlp.vocab)
|
||
phrasematcher.add('OBAMA', None, nlp(u"Barack Obama"))
|
||
|
||
p
|
||
| Patterns can now be added to the matcher by calling
|
||
| #[+api("matcher#add") #[code matcher.add()]] with a match ID, an optional
|
||
| callback function to be invoked on each match, and one or more patterns.
|
||
| This allows you to write powerful, pattern-specific logic using only one
|
||
| matcher. For example, you might only want to merge some entity types,
|
||
| and set custom flags for other matched patterns. The new
|
||
| #[+api("phrasematcher") #[code PhraseMatcher]] lets you efficiently
|
||
| match very large terminology lists using #[code Doc] objects as match
|
||
| patterns.
|
||
|
||
+infobox
|
||
| #[+label-inline API:] #[+api("matcher") #[code Matcher]],
|
||
| #[+api("phrasematcher") #[code PhraseMatcher]]
|
||
| #[+label-inline Usage:]
|
||
| #[+a("/usage/linguistic-features#rule-based-matching") Rule-based matching]
|