* OrigAnnot class instead of gold.orig_annot list of zipped tuples
* from_orig to replace from_annot_tuples
* rename to RawAnnot
* some unit tests for GoldParse creation and internal format
* removing orig_annot and switching to lists instead of tuple
* rewriting tuples to use RawAnnot (+ debug statements, WIP)
* fix pop() changing the data
* small fixes
* pop-append fixes
* return RawAnnot for existing GoldParse to have uniform interface
* clean up imports
* fix merge_sents
* add unit test for 4402 with new structure (not working yet)
* introduce DocAnnot
* typo fixes
* add unit test for merge_sents
* rename from_orig to from_raw
* fixing unit tests
* fix nn parser
* read_annots to produce text, doc_annot pairs
* _make_golds fix
* rename golds_to_gold_annots
* small fixes
* fix encoding
* have golds_to_gold_annots use DocAnnot
* missed a spot
* merge_sents as function in DocAnnot
* allow specifying only part of the token-level annotations
* refactor with Example class + underlying dicts
* pipeline components to work with Example objects (wip)
* input checking
* fix yielding
* fix calls to update
* small fixes
* fix scorer unit test with new format
* fix kwargs order
* fixes for ud and conllu scripts
* fix reading data for conllu script
* add in proper errors (not fixed numbering yet to avoid merge conflicts)
* fixing few more small bugs
* fix EL script
The new spacy pretrain command implemented BERT/ULMFit/etc-like transfer learning, using our Language Modelling with Approximate Outputs version of BERT's cloze task. Pretraining is convenient, but in some ways it's a bit of a strange solution. All we're doing is initialising the weights. At the same time, we're putting a lot of work into our optimisation so that it's less sensitive to initial conditions, and more likely to find good optima. I discuss this a bit in the pseudo-rehearsal blog post: https://explosion.ai/blog/pseudo-rehearsal-catastrophic-forgetting
Support semi-supervised learning in spacy train
One obvious way to improve these pretraining methods is to do multi-task learning, instead of just transfer learning. This has been shown to work very well: https://arxiv.org/pdf/1809.08370.pdf . This patch makes it easy to do this sort of thing.
Add a new argument to spacy train, --raw-text. This takes a jsonl file with unlabelled data that can be used in arbitrary ways to do semi-supervised learning.
Add a new method to the Language class and to pipeline components, .rehearse(). This is like .update(), but doesn't expect GoldParse objects. It takes a batch of Doc objects, and performs an update on some semi-supervised objective.
Move the BERT-LMAO objective out from spacy/cli/pretrain.py into spacy/_ml.py, so we can create a new pipeline component, ClozeMultitask. This can be specified as a parser or NER multitask in the spacy train command. Example usage:
python -m spacy train en ./tmp ~/data/en-core-web/train/nw.json ~/data/en-core-web/dev/nw.json --pipeline parser --raw-textt ~/data/unlabelled/reddit-100k.jsonl --vectors en_vectors_web_lg --parser-multitasks cloze
Implement rehearsal methods for pipeline components
The new --raw-text argument and nlp.rehearse() method also gives us a good place to implement the the idea in the pseudo-rehearsal blog post in the parser. This works as follows:
Add a new nlp.resume_training() method. This allocates copies of pre-trained models in the pipeline, setting things up for the rehearsal updates. It also returns an optimizer object. This also greatly reduces confusion around the nlp.begin_training() method, which randomises the weights, making it not suitable for adding new labels or otherwise fine-tuning a pre-trained model.
Implement rehearsal updates on the Parser class, making it available for the dependency parser and NER. During rehearsal, the initial model is used to supervise the model being trained. The current model is asked to match the predictions of the initial model on some data. This minimises catastrophic forgetting, by keeping the model's predictions close to the original. See the blog post for details.
Implement rehearsal updates for tagger
Implement rehearsal updates for text categoriz
Sofie Van Landeghem