Commit Graph

21 Commits

Author SHA1 Message Date
Matthew Honnibal
8f2a6367e9 Fix usage of PyTorch BiLSTM in ud_train 2018-09-13 22:54:59 +00:00
Matthew Honnibal
445b81ce3f Support bilstm_depth argument in ud-train 2018-09-13 19:30:22 +02:00
Matthew Honnibal
3eb9f3e2b8 Fix defaults for ud-train 2018-09-13 18:05:48 +02:00
Matthew Honnibal
59cf533879 Improve ud-train script. Make config optional 2018-09-13 14:24:08 +02:00
Matthew Honnibal
da7650e84b Fix maximum doc length in ud_train script 2018-09-13 14:10:25 +02:00
Matthew Honnibal
4d2d7d5866 Fix new feature flags 2018-08-27 02:12:39 +02:00
Matthew Honnibal
9c33d4d1df Add more hyper-parameters to spacy ud-train
* subword_features: Controls whether subword features are used in the
word embeddings. True by default (specifically, prefix, suffix and word
shape). Should be set to False for languages like Chinese and Japanese.

* conv_depth: Depth of the convolutional layers. Defaults to 4.
2018-08-27 01:48:46 +02:00
Matthew Honnibal
74d5c625b3 Use rising beam update prob 2018-05-16 20:11:59 +02:00
Matthew Honnibal
dc1a479fbd Merge branch 'develop' into feature/refactor-parser 2018-05-15 18:39:21 +02:00
Matthew Honnibal
fc4dd49b77 Support oracle segmentation in ud-train CLI command 2018-05-08 13:47:45 +02:00
Matthew Honnibal
eddc0e0c74 Set gold.sent_starts in ud_train 2018-05-07 15:52:47 +02:00
Matthew Honnibal
17af6aa3a4 Update ud_train script 2018-04-29 15:49:32 +02:00
Matthew Honnibal
8bbd26579c Support GPU in UD training script 2018-03-27 09:53:35 +00:00
Matthew Honnibal
85717f570c Merge branch 'master' of https://github.com/explosion/spaCy 2018-03-23 20:30:42 +01:00
Matthew Honnibal
8902754f0b Fix vector loading for ud_train 2018-03-23 20:30:00 +01:00
Xiaoquan Kong
a71b99d7ff bugfix for global-variable-change-in-runtime related issue (#2135)
* Bugfix: setting pollution from spacy/cli/ud_train.py to whole package

* Add contributor agreement of howl-anderson
2018-03-23 11:36:38 +01:00
Matthew Honnibal
bede11b67c
Improve label management in parser and NER (#2108)
This patch does a few smallish things that tighten up the training workflow a little, and allow memory use during training to be reduced by letting the GoldCorpus stream data properly.

Previously, the parser and entity recognizer read and saved labels as lists, with extra labels noted separately. Lists were used becaue ordering is very important, to ensure that the label-to-class mapping is stable.

We now manage labels as nested dictionaries, first keyed by the action, and then keyed by the label. Values are frequencies. The trick is, how do we save new labels? We need to make sure we iterate over these in the same order they're added. Otherwise, we'll get different class IDs, and the model's predictions won't make sense.

To allow stable sorting, we map the new labels to negative values. If we have two new labels, they'll be noted as having "frequency" -1 and -2. The next new label will then have "frequency" -3. When we sort by (frequency, label), we then get a stable sort.

Storing frequencies then allows us to make the next nice improvement. Previously we had to iterate over the whole training set, to pre-process it for the deprojectivisation. This led to storing the whole training set in memory. This was most of the required memory during training.

To prevent this, we now store the frequencies as we stream in the data, and deprojectivize as we go. Once we've built the frequencies, we can then apply a frequency cut-off when we decide how many classes to make.

Finally, to allow proper data streaming, we also have to have some way of shuffling the iterator. This is awkward if the training files have multiple documents in them. To solve this, the GoldCorpus class now writes the training data to disk in msgpack files, one per document. We can then shuffle the data by shuffling the paths.

This is a squash merge, as I made a lot of very small commits. Individual commit messages below.

* Simplify label management for TransitionSystem and its subclasses

* Fix serialization for new label handling format in parser

* Simplify and improve GoldCorpus class. Reduce memory use, write to temp dir

* Set actions in transition system

* Require thinc 6.11.1.dev4

* Fix error in parser init

* Add unicode declaration

* Fix unicode declaration

* Update textcat test

* Try to get model training on less memory

* Print json loc for now

* Try rapidjson to reduce memory use

* Remove rapidjson requirement

* Try rapidjson for reduced mem usage

* Handle None heads when projectivising

* Stream json docs

* Fix train script

* Handle projectivity in GoldParse

* Fix projectivity handling

* Add minibatch_by_words util from ud_train

* Minibatch by number of words in spacy.cli.train

* Move minibatch_by_words util to spacy.util

* Fix label handling

* More hacking at label management in parser

* Fix encoding in msgpack serialization in GoldParse

* Adjust batch sizes in parser training

* Fix minibatch_by_words

* Add merge_subtokens function to pipeline.pyx

* Register merge_subtokens factory

* Restore use of msgpack tmp directory

* Use minibatch-by-words in train

* Handle retokenization in scorer

* Change back-off approach for missing labels. Use 'dep' label

* Update NER for new label management

* Set NER tags for over-segmented words

* Fix label alignment in gold

* Fix label back-off for infrequent labels

* Fix int type in labels dict key

* Fix int type in labels dict key

* Update feature definition for 8 feature set

* Update ud-train script for new label stuff

* Fix json streamer

* Print the line number if conll eval fails

* Update children and sentence boundaries after deprojectivisation

* Export set_children_from_heads from doc.pxd

* Render parses during UD training

* Remove print statement

* Require thinc 6.11.1.dev6. Try adding wheel as install_requires

* Set different dev version, to flush pip cache

* Update thinc version

* Update GoldCorpus docs

* Remove print statements

* Fix formatting and links [ci skip]
2018-03-19 02:58:08 +01:00
Matthew Honnibal
d7ce6527fb Use increasing batch sizes in ud-train 2018-03-14 20:15:28 +01:00
Matthew Honnibal
5dddb30e5b Fix ud-train script 2018-03-11 01:26:45 +01:00
Matthew Honnibal
2cab4d6517 Remove use of attr module in ud_train 2018-03-11 00:59:39 +01:00
Matthew Honnibal
3478ea76d1 Add ud_train and ud_evaluate CLI commands 2018-03-10 23:41:55 +01:00