Matthew Honnibal 375f0dc529 💫 Make TextCategorizer default to a simpler, GPU-friendly model (#3038) ... Currently the TextCategorizer defaults to a fairly complicated model, designed partly around the active learning requirements of Prodigy. The model's a bit slow, and not very GPU-friendly. This patch implements a straightforward CNN model that still performs pretty well. The replacement model also makes it easy to use the LMAO pretraining, since most of the parameters are in the CNN. The replacement model has a flag to specify whether labels are mutually exclusive, which defaults to True. This has been a common problem with the text classifier. We'll also now be able to support adding labels to pretrained models again. Resolves #2934, #2756, #1798, #1748.		2018-12-10 14:37:39 +01:00
..
conllu.py	Remove unused cytoolz / itertools imports	2018-12-03 02:12:07 +01:00
ner_multitask_objective.py	Auto-format examples	2018-12-02 04:26:26 +01:00
pretrain_textcat.py	Auto-format examples	2018-12-02 04:26:26 +01:00
train_intent_parser.py	Auto-format examples	2018-12-02 04:26:26 +01:00
train_ner.py	Auto-format examples	2018-12-02 04:26:26 +01:00
train_new_entity_type.py	Auto-format examples	2018-12-02 04:26:26 +01:00
train_parser.py	Auto-format examples	2018-12-02 04:26:26 +01:00
train_tagger.py	Auto-format examples	2018-12-02 04:26:26 +01:00
train_textcat.py	💫 Make TextCategorizer default to a simpler, GPU-friendly model (#3038)	2018-12-10 14:37:39 +01:00
training-data.json	Use better training data JSON example	2017-10-24 16:00:56 +02:00
vocab-data.jsonl	Use even smaller examle size	2017-10-30 19:46:45 +01:00