145 Commits

Author	SHA1	Message	Date
Matthew Honnibal	8075726838	Restore vector usage in models	2017-10-31 19:21:17 +01:00
Matthew Honnibal	cb5217012f	Fix vector remapping	2017-10-31 11:40:46 +01:00
Matthew Honnibal	368fdb389a	WIP on refactoring and fixing vectors	2017-10-31 02:00:26 +01:00
Matthew Honnibal	3b91097321	Whitespace	2017-10-28 17:05:11 +00:00
Matthew Honnibal	6ef72864fa	Improve initialization for hidden layers	2017-10-28 17:05:01 +00:00
Matthew Honnibal	df4803cc6d	Add learned missing values for parser	2017-10-28 16:45:14 +00:00
Matthew Honnibal	64e4ff7c4b	Merge 'tidy-up' changes into branch. Resolve conflicts	2017-10-28 13:16:06 +02:00
Explosion Bot	b22e42af7f	Merge changes to parser and _ml	2017-10-28 11:52:10 +02:00
ines	d96e72f656	Tidy up rest	2017-10-27 21:07:59 +02:00
ines	e33b7e0b3c	Tidy up parser and ML	2017-10-27 14:39:30 +02:00
Matthew Honnibal	531142a933	Merge remote-tracking branch 'origin/develop' into feature/better-parser	2017-10-27 12:34:48 +00:00
Matthew Honnibal	c9987cf131	Avoid use of numpy.tensordot	2017-10-27 10:18:36 +00:00
Matthew Honnibal	f6fef30adc	Remove dead code from spacy._ml	2017-10-27 10:16:41 +00:00
ines	4eb5bd02e7	Update textcat pre-processing after to_array change	2017-10-27 00:32:12 +02:00
Matthew Honnibal	35977bdbb9	Update better-parser branch with develop	2017-10-26 00:55:53 +00:00
Matthew Honnibal	075e8118ea	Update from develop	2017-10-25 12:45:21 +02:00
ines	0b1dcbac14	Remove unused function	2017-10-25 12:08:46 +02:00
Matthew Honnibal	3faf9189a2	Make parser hidden shape consistent even if maxout==1	2017-10-20 16:23:31 +02:00
Matthew Honnibal	b101736555	Fix precomputed layer	2017-10-20 12:14:52 +02:00
Matthew Honnibal	64658e02e5	Implement fancier initialisation for precomputed layer	2017-10-20 03:07:45 +02:00
Matthew Honnibal	a17a1b60c7	Clean up redundant PrecomputableMaxouts class	2017-10-19 20:26:37 +02:00
Matthew Honnibal	b00d0a2c97	Fix bias in parser	2017-10-19 18:42:11 +02:00
Matthew Honnibal	03a215c5fd	Make PrecomputableAffines work	2017-10-19 13:44:49 +02:00
Matthew Honnibal	76fe24f44d	Improve embedding defaults	2017-10-11 09:44:17 +02:00
Matthew Honnibal	b2b8506f2c	Remove whitespace	2017-10-09 03:35:57 +02:00
Matthew Honnibal	d163115e91	Add non-linearity after history features	2017-10-07 21:00:43 -05:00
Matthew Honnibal	5c750a9c2f	Reserve 0 for 'missing' in history features	2017-10-06 06:10:13 -05:00
Matthew Honnibal	fbba7c517e	Pass dropout through to embed tables	2017-10-06 06:09:18 -05:00
Matthew Honnibal	3db0a32fd6	Fix dropout for history features	2017-10-05 22:21:30 -05:00
Matthew Honnibal	fc06b0a333	Fix training when hist_size==0	2017-10-05 21:52:28 -05:00
Matthew Honnibal	dcdfa071aa	Disable LayerNorm hack	2017-10-04 20:06:52 -05:00
Matthew Honnibal	bfabc333be	Merge remote-tracking branch 'origin/develop' into feature/parser-history-model	2017-10-04 20:00:36 -05:00
Matthew Honnibal	92066b04d6	Fix Embed and HistoryFeatures	2017-10-04 19:55:34 -05:00
Matthew Honnibal	bd8e84998a	Add nO attribute to TextCategorizer model	2017-10-04 16:07:30 +02:00
Matthew Honnibal	f8a0614527	Improve textcat model slightly	2017-10-04 15:15:53 +02:00
Matthew Honnibal	39798b0172	Uncomment layernorm adjustment hack	2017-10-04 15:12:09 +02:00
Matthew Honnibal	774f5732bd	Fix dimensionality of textcat when no vectors available	2017-10-04 14:55:15 +02:00
Matthew Honnibal	af75b74208	Unset LayerNorm backwards compat hack	2017-10-03 20:47:10 -05:00
Matthew Honnibal	246612cb53	Merge remote-tracking branch 'origin/develop' into feature/parser-history-model	2017-10-03 16:56:42 -05:00
Matthew Honnibal	5cbefcba17	Set backwards compatibility flag	2017-10-03 20:29:58 +02:00
Matthew Honnibal	5454b20cd7	Update thinc imports for 6.9	2017-10-03 20:07:17 +02:00
Matthew Honnibal	e514d6aa0a	Import thinc modules more explicitly, to avoid cycles	2017-10-03 18:49:25 +02:00
Matthew Honnibal	b770f4e108	Fix embed class in history features	2017-10-03 13:26:55 +02:00
Matthew Honnibal	6aa6a5bc25	Add a layer type for history features	2017-10-03 12:43:09 +02:00
Matthew Honnibal	f6330d69e6	Default embed size to 7000	2017-09-28 08:07:41 -05:00
Matthew Honnibal	1a37a2c0a0	Update training defaults	2017-09-27 11:48:07 -05:00
Matthew Honnibal	e34e70673f	Allow tagger models to be built with pre-defined tok2vec layer	2017-09-26 05:51:52 -05:00
Matthew Honnibal	63bd87508d	Don't use iterated convolutions	2017-09-23 04:39:17 -05:00
Matthew Honnibal	4348c479fc	Merge pre-trained vectors and noshare patches	2017-09-22 20:07:28 -05:00
Matthew Honnibal	4bd6a12b1f	Fix Tok2Vec	2017-09-23 02:58:54 +02:00