explosion/spaCy
1
1
Fork 0
mirror of https://github.com/explosion/spaCy.git synced 2025-01-26 09:14:32 +03:00
Code Issues Packages Projects Releases Wiki Activity

Add intro narrative for v2

Browse Source
This commit is contained in:
Matthew Honnibal 2017-06-04 15:10:37 +02:00
parent a66cf24ee8
commit 23fd6b1782
1 changed files with 40 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

41
website/docs/usage/v2.jade
View File

@ -3,7 +3,46 @@
include ../../_includes/_mixins
p
| We're very excited to finally introduce spaCy v2.0.
| We're very excited to finally introduce spaCy v2.0. This release features
| entirely new deep learning-powered models for spaCy's tagger, parser and
| entity recognizer. The new models are #[strong 20x smaller] than the linear
| models that have powered spaCy until now: from 300mb to only 14mb. Speed
| and accuracy are currently comparable to the 1.x models: speed on CPU is
| slightly lower, while accuracy is slightly higher. We expect performance to
| improve quickly between now and the release date, as we run more experiments
| and optimize the implementation.
p
| The main usability improvements you'll notice in spaCy 2 are around the
| defining, training and loading your own models and components. The new neural
| network models make it much easier to train a model from scratch, or update
| an existing model with a few examples. In v1, the statistical models depended
| on the state of the vocab. If you taught the model a new word, you would have
| to save and load a lot of data -- otherwise the model wouldn't correctly
| recall the features of your new example. That's no longer the case. Due to some
| clever use of hashing, the statistical models never change size, even as they
| learn new vocabulary items. The whole pipeline is also now fully differentiable,
| so even if you don't have explicitly annotated data, you can update spaCy using
| all the latest deep learning tricks: adversarial training, noise contrastive
| estimation, reinforcement learning, etc.
p
| Finally, we've made several usability improvements that are particularly helpful
| for production deployments. spaCy 2 now fully supports the Pickle protocol,
| making it easy to use spaCy with Apache Spark. The string-to-integer mapping is
| no longer stateful, making it easy to reconcile annotations made in different
| processes. Models are smaller and use less memory, and the APIs for serialization
| are now much more consistent.
p
| Because we'e made so many architectural changes to the library, we've tried to
| keep breaking changes to a minimum. A lot of projects follow the philosophy that
| if you're going to break anything, you may as well break everything. We think
| migration is easier if there's a logic to what's changed. We've therefore followed
| a policy of avoiding breaking changes to the #[code Doc], #[code Span] and #[code Token]
| objects. This way, you can focus on only migrating the code that does training, loading
| and serialisation --- in other words, code that works with the #[code nlp] object directly.
| Code that uses the annotations should continue to work.
p
| On this page, you'll find a summary of the #[+a("#features") new features],