Sofie Van Landeghem
cd70c3cb79
Fixing pretrain ( #7342 )
...
* initialize NLP with train corpus
* add more pretraining tests
* more tests
* function to fetch tok2vec layer for pretraining
* clarify parameter name
* test different objectives
* formatting
* fix check for static vectors when using vectors objective
* clarify docs
* logger statement
* fix init_tok2vec and proc.initialize order
* test training after pretraining
* add init_config tests for pretraining
* pop pretraining block to avoid config validation errors
* custom errors
2021-03-09 14:01:13 +11:00
Adriane Boyd
97bcf2ae3a
Fix patience for identical scores ( #7250 )
...
* Fix patience for identical scores
Fix training patience so that the earliest best step is chosen for
identical max scores.
* Restore break, remove print
* Explicitly define best_step for clarity
2021-03-06 18:42:14 +11:00
Ines Montani
ea555b03e0
Merge pull request #7255 from adrianeboyd/bugfix/extraneous-tok2vec
...
Omit unused tok2vec/transformer components
2021-03-03 23:15:06 +11:00
svlandeg
d900c55061
consistently use registry as callable
2021-03-02 17:56:28 +01:00
Adriane Boyd
8a4200d4e9
Omit unused tok2vec/transformer components
...
Omit unused tok2vec/transformer components in quickstart template.
2021-03-02 15:53:30 +01:00
Sofie Van Landeghem
212f0e779e
Support doc.spans in Example.from_dict ( #7197 )
...
* add support for spans in Example.from_dict
* add unit tests
* update error to E879
2021-03-03 01:12:54 +11:00
Adriane Boyd
fb98862337
Add hint for --gpu-id to CLI device info ( #7234 )
...
* Add hint for --gpu-id to CLI device info
If the user has `cupy` and an available GPU, add a hint about using
`--gpu-id 0` to the CLI output.
* Undo change to original CPU message
2021-03-03 01:11:18 +11:00
Ines Montani
635ae55b74
Merge pull request #7237 from adrianeboyd/bugfix/is-cython-func-7224
2021-03-03 00:05:16 +11:00
Adriane Boyd
0efb7413f9
Use make_tempdir instead
2021-03-01 17:54:14 +01:00
Adriane Boyd
e9f7f9a4bc
Fix is_cython_func for additional imported code
...
* Fix `is_cython_func` for imported code loaded under `python_code`
module name
* Add `make_named_tempfile` context manager to test utils to test
loading of imported code
* Add test for validation of `initialize` params in custom module
2021-03-01 16:37:39 +01:00
Sofie Van Landeghem
dd99872bb0
Fix spans weak ref in doc copy ( #7225 )
...
* failing unit test
* ensure that doc.spans refers to the copied doc, not the old
* add type info
2021-02-28 12:32:48 +11:00
Ines Montani
408b94887a
Merge pull request #7207 from adrianeboyd/docs/get-noun-chunks [ci skip]
...
Extend docs related to Vocab.get_noun_chunks
2021-02-27 11:51:08 +11:00
Ines Montani
dc46fa078f
Merge pull request #7220 from svlandeg/docs/has_annotation [ci skip]
...
has_annotation docs fix
2021-02-27 11:50:34 +11:00
Ines Montani
0dbc2a1b16
Merge pull request #7222 from adrianeboyd/bugfix/quickstart-recs-bg-bn
...
Fix formatting in bg/bn quickstart recs
2021-02-27 11:50:02 +11:00
svlandeg
2010219a7f
import wandb failure - UX
2021-02-26 18:00:39 +01:00
Adriane Boyd
ee7bb0b393
Fix formatting in bg/bn quickstart recs
2021-02-26 17:08:37 +01:00
svlandeg
248339039e
fix type in docs
2021-02-26 14:27:10 +01:00
Adriane Boyd
e43d43db32
Allow sourcing disabled components ( #7215 )
...
Check `component_names` instead of `pipe_names` to allow sourcing
disabled components.
2021-02-26 13:50:56 +01:00
Adriane Boyd
10c930cc96
Re-refactor Sentencizer with Pipe API ( #7176 )
...
Reapply the refactoring (#4721 ) so that `Sentencizer` uses the faster
`predict` and `set_annotations` for both `__call__` and `pipe`.
2021-02-26 09:48:14 +01:00
Adriane Boyd
6a37f343d5
Extend docs related to Vocab.get_noun_chunks
2021-02-25 16:38:21 +01:00
Ines Montani
592678fb7d
Merge pull request #7073 from adrianeboyd/feature/logger-level-in-formatter
...
Add time and level to default logging formatter
2021-02-24 22:40:46 +11:00
Sofie Van Landeghem
0563cd73d6
Fix SpanGroup import ( #7182 )
...
* import SpanGroup from tokens module
* revert edits from different PR
* add to __all__
2021-02-24 21:06:16 +11:00
Sofie Van Landeghem
b92f81d5da
fix NEL config and IO, and n_sents functionality ( #7100 )
...
* fix NEL config and IO, and n_sents functionality
* add docs
* fix test
2021-02-22 14:49:52 +11:00
Sofie Van Landeghem
113e8d082b
only evaluate named entities for NEL if there is a corresponding gold span ( #7074 )
2021-02-22 11:06:50 +11:00
Adriane Boyd
264862c67a
Fix Ukrainian lemmatizer init ( #7127 )
...
Fix class variable and init for `UkrainianLemmatizer` so that it loads
the `uk` dictionaries rather than having the parent `RussianLemmatizer`
override with the `ru` settings.
2021-02-22 11:05:08 +11:00
Sofie Van Landeghem
ba5a50f62b
NEL docs & UX ( #7129 )
...
* EL set_kb docs fix
* custom warning for set_kb mistake
2021-02-22 11:04:22 +11:00
Boian Tzonev
cca8651fc8
Bulgarian tokenizer exceptions ( #7114 )
...
* [Bulgarian] Add tokenizer exceptions and like_num for Bulgarian
* [Bulgarian] Add tokenizer exceptions and like_num for Bulgarian
2021-02-19 19:19:19 +01:00
Sofie Van Landeghem
709c9e75af
span.ent only returns first sentence ( #7084 )
...
* return first sentence when span contains sentence boundary
* docs fix
* small fixes
* cleanup
2021-02-19 23:02:38 +11:00
Adriane Boyd
30e1a89aeb
Fix displacy output in evaluate CLI ( #7122 )
...
Now that `nlp.evaluate()` does not modify the examples, rerun the
pipeline on the (limited) texts in order to provide the predicted
annotation in the displacy output option.
2021-02-19 23:01:20 +11:00
Adriane Boyd
4188beda87
Fix conll converter option ( #7071 )
...
Map `conll` to the NER converter, not the `CoNLL-U` converter.
2021-02-18 10:22:41 +01:00
Adriane Boyd
a3293efc48
Add time and level to default logging formatter
2021-02-15 14:19:20 +01:00
Ines Montani
1e3a326e53
Change Dutch transformer recommendation [ci skip]
...
https://github.com/explosion/spaCy/discussions/6529#discussioncomment-366620
2021-02-14 15:30:16 +11:00
Ines Montani
f4f46b617f
Preserve sourced components in fill-config ( fixes #7055 ) ( #7058 )
2021-02-14 14:02:14 +11:00
Matthew Honnibal
0fb8d437c0
Fix sentence fragments bug ( #7056 , #7035 ) ( #7057 )
...
* Add test for #7035
* Update test for issue 7056
* Fix test
* Fix transitions method used in testing
* Fix state eol detection when rebuffer
* Clean up redundant fix
2021-02-14 13:38:13 +11:00
Ines Montani
660642902a
Increment version [ci skip]
2021-02-14 13:36:13 +11:00
Matthew Honnibal
b31471b5b8
Set version to v3.0.2
2021-02-13 23:50:00 +11:00
Ines Montani
9ba715ed16
Tidy up and auto-format
2021-02-13 12:55:56 +11:00
Ines Montani
34ee0fbd70
Merge pull request #7011 from Shumie82/master
2021-02-13 12:30:42 +11:00
Ines Montani
e583050547
Merge pull request #7039 from svlandeg/debug
2021-02-13 11:53:41 +11:00
Ines Montani
6c450decfc
Fix punctuation settings and add to initialize tests
2021-02-13 11:51:21 +11:00
Ines Montani
f4712a634e
Merge pull request #7046 from adrianeboyd/bugfix/vocab-pickle-noun-chunks-6891
...
Include noun chunks method when pickling Vocab
2021-02-13 11:43:03 +11:00
Adriane Boyd
0ee2ae86bf
Update trf quickstart recommendations
...
Add/update trf recommendations for Bengali, Hindi, Sinhala, and Tamil
based on #7044 .
2021-02-12 15:55:17 +01:00
svlandeg
03b4ec7d7f
fix typo
2021-02-12 14:30:16 +01:00
Adriane Boyd
5e47a54d29
Include noun chunks method when pickling Vocab
2021-02-12 13:27:46 +01:00
svlandeg
aa3ad8825d
loop instead of any
2021-02-12 13:14:30 +01:00
svlandeg
278e9eaa14
remove ner
2021-02-11 21:08:04 +01:00
svlandeg
ebeedfc70b
regression test for 7029
2021-02-11 20:56:48 +01:00
svlandeg
a52d466bfc
any instead of all
2021-02-11 20:50:55 +01:00
Shumi
4e514f1ea8
Update stop_words.py
...
I have deleted line 1 to 5 and the statement print(STOP_WORDS)
2021-02-11 21:30:34 +02:00
Shumi
0d57e84b7b
Update lex_attrs.py
...
I have removed line 1 to 4
2021-02-11 21:28:23 +02:00