README.md

Sequence-Tagging

Dataset

The introduction to dataset: README.md.

Pre-trained Word Embedding & Language Model

• Convert your BERT from tensorflow to pytorch: scripts/bert_convert.sh.
• Train your GloVe using stanfordnlp/GloVe: scripts/glove_train.sh.

Run

Train:

allennlp train configs/bilstm_crf.jsonnet -s outputs/bilstm_crf -f --include-package hanzo

Predict:

allennlp predict outputs/bilstm_crf/model.tar.gz resources/datagrand/test.txt \
                 --include-package hanzo \
                 --predictor data_grand \
                 --use-dataset-reader \
                 --silent \
                 --output-file result.txt