├── .gitignore ├── .gitmodules ├── LICENSE ├── README.md ├── bert_entity ├── __init__.py ├── data_loader_conll.py ├── data_loader_wiki.py ├── metrics.py ├── misc.py ├── model.py ├── model_conll.py ├── pipeline_job.py ├── preprocess_all.py ├── preprocessing │ ├── __init__.py │ ├── collect_mention_entity_counts.py │ ├── create_disambiguation_dict.py │ ├── create_integerized_aida_conll_training.py │ ├── create_integerized_wiki_training.py │ ├── create_keyword_matcher.py │ ├── create_redirects.py │ ├── create_resolve_to_wiki_dicts.py │ ├── create_wiki_training_data.py │ ├── download_data.py │ ├── postprocess_mention_entity_counts.py │ ├── preprocess_aida_conll_data.py │ └── wikiextractor.py ├── train.py ├── train_util.py └── vocab.py ├── config ├── conll2019__preprocess.yaml ├── conll2019__train_on_aida_conll.yaml ├── conll2019__train_on_wiki.yaml ├── dummy__preprocess.yaml ├── dummy__train_on_aida_conll.yaml └── dummy__train_on_wiki.yaml ├── docs ├── Bert-Entity.png └── preprocessing.png ├── downstream_experiments ├── ensemble_bert_modeling.py ├── fairseq_patch_01.patch ├── fairseq_patch_02.patch ├── prepare_fairseq.sh ├── run_bert_glue_classifier.py ├── run_bert_glue_classifier_ensemble.py ├── run_bert_squad.py ├── run_bert_squad_ensemble.py ├── run_bert_swag_ensemble.py └── run_fairseq_bert_ensemble.sh ├── requirements.txt └── setup_paths /.gitignore: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/samuelbroscheit/entity_knowledge_in_bert/HEAD/.gitignore -------------------------------------------------------------------------------- /.gitmodules: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/samuelbroscheit/entity_knowledge_in_bert/HEAD/.gitmodules -------------------------------------------------------------------------------- /LICENSE: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/samuelbroscheit/entity_knowledge_in_bert/HEAD/LICENSE -------------------------------------------------------------------------------- /README.md: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/samuelbroscheit/entity_knowledge_in_bert/HEAD/README.md -------------------------------------------------------------------------------- /bert_entity/__init__.py: -------------------------------------------------------------------------------- 1 | -------------------------------------------------------------------------------- /bert_entity/data_loader_conll.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/samuelbroscheit/entity_knowledge_in_bert/HEAD/bert_entity/data_loader_conll.py -------------------------------------------------------------------------------- /bert_entity/data_loader_wiki.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/samuelbroscheit/entity_knowledge_in_bert/HEAD/bert_entity/data_loader_wiki.py -------------------------------------------------------------------------------- /bert_entity/metrics.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/samuelbroscheit/entity_knowledge_in_bert/HEAD/bert_entity/metrics.py -------------------------------------------------------------------------------- /bert_entity/misc.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/samuelbroscheit/entity_knowledge_in_bert/HEAD/bert_entity/misc.py -------------------------------------------------------------------------------- /bert_entity/model.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/samuelbroscheit/entity_knowledge_in_bert/HEAD/bert_entity/model.py -------------------------------------------------------------------------------- /bert_entity/model_conll.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/samuelbroscheit/entity_knowledge_in_bert/HEAD/bert_entity/model_conll.py -------------------------------------------------------------------------------- /bert_entity/pipeline_job.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/samuelbroscheit/entity_knowledge_in_bert/HEAD/bert_entity/pipeline_job.py -------------------------------------------------------------------------------- /bert_entity/preprocess_all.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/samuelbroscheit/entity_knowledge_in_bert/HEAD/bert_entity/preprocess_all.py -------------------------------------------------------------------------------- /bert_entity/preprocessing/__init__.py: -------------------------------------------------------------------------------- 1 | -------------------------------------------------------------------------------- /bert_entity/preprocessing/collect_mention_entity_counts.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/samuelbroscheit/entity_knowledge_in_bert/HEAD/bert_entity/preprocessing/collect_mention_entity_counts.py -------------------------------------------------------------------------------- /bert_entity/preprocessing/create_disambiguation_dict.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/samuelbroscheit/entity_knowledge_in_bert/HEAD/bert_entity/preprocessing/create_disambiguation_dict.py -------------------------------------------------------------------------------- /bert_entity/preprocessing/create_integerized_aida_conll_training.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/samuelbroscheit/entity_knowledge_in_bert/HEAD/bert_entity/preprocessing/create_integerized_aida_conll_training.py -------------------------------------------------------------------------------- /bert_entity/preprocessing/create_integerized_wiki_training.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/samuelbroscheit/entity_knowledge_in_bert/HEAD/bert_entity/preprocessing/create_integerized_wiki_training.py -------------------------------------------------------------------------------- /bert_entity/preprocessing/create_keyword_matcher.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/samuelbroscheit/entity_knowledge_in_bert/HEAD/bert_entity/preprocessing/create_keyword_matcher.py -------------------------------------------------------------------------------- /bert_entity/preprocessing/create_redirects.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/samuelbroscheit/entity_knowledge_in_bert/HEAD/bert_entity/preprocessing/create_redirects.py -------------------------------------------------------------------------------- /bert_entity/preprocessing/create_resolve_to_wiki_dicts.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/samuelbroscheit/entity_knowledge_in_bert/HEAD/bert_entity/preprocessing/create_resolve_to_wiki_dicts.py -------------------------------------------------------------------------------- /bert_entity/preprocessing/create_wiki_training_data.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/samuelbroscheit/entity_knowledge_in_bert/HEAD/bert_entity/preprocessing/create_wiki_training_data.py -------------------------------------------------------------------------------- /bert_entity/preprocessing/download_data.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/samuelbroscheit/entity_knowledge_in_bert/HEAD/bert_entity/preprocessing/download_data.py -------------------------------------------------------------------------------- /bert_entity/preprocessing/postprocess_mention_entity_counts.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/samuelbroscheit/entity_knowledge_in_bert/HEAD/bert_entity/preprocessing/postprocess_mention_entity_counts.py -------------------------------------------------------------------------------- /bert_entity/preprocessing/preprocess_aida_conll_data.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/samuelbroscheit/entity_knowledge_in_bert/HEAD/bert_entity/preprocessing/preprocess_aida_conll_data.py -------------------------------------------------------------------------------- /bert_entity/preprocessing/wikiextractor.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/samuelbroscheit/entity_knowledge_in_bert/HEAD/bert_entity/preprocessing/wikiextractor.py -------------------------------------------------------------------------------- /bert_entity/train.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/samuelbroscheit/entity_knowledge_in_bert/HEAD/bert_entity/train.py -------------------------------------------------------------------------------- /bert_entity/train_util.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/samuelbroscheit/entity_knowledge_in_bert/HEAD/bert_entity/train_util.py -------------------------------------------------------------------------------- /bert_entity/vocab.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/samuelbroscheit/entity_knowledge_in_bert/HEAD/bert_entity/vocab.py -------------------------------------------------------------------------------- /config/conll2019__preprocess.yaml: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/samuelbroscheit/entity_knowledge_in_bert/HEAD/config/conll2019__preprocess.yaml -------------------------------------------------------------------------------- /config/conll2019__train_on_aida_conll.yaml: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/samuelbroscheit/entity_knowledge_in_bert/HEAD/config/conll2019__train_on_aida_conll.yaml -------------------------------------------------------------------------------- /config/conll2019__train_on_wiki.yaml: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/samuelbroscheit/entity_knowledge_in_bert/HEAD/config/conll2019__train_on_wiki.yaml -------------------------------------------------------------------------------- /config/dummy__preprocess.yaml: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/samuelbroscheit/entity_knowledge_in_bert/HEAD/config/dummy__preprocess.yaml -------------------------------------------------------------------------------- /config/dummy__train_on_aida_conll.yaml: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/samuelbroscheit/entity_knowledge_in_bert/HEAD/config/dummy__train_on_aida_conll.yaml -------------------------------------------------------------------------------- /config/dummy__train_on_wiki.yaml: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/samuelbroscheit/entity_knowledge_in_bert/HEAD/config/dummy__train_on_wiki.yaml -------------------------------------------------------------------------------- /docs/Bert-Entity.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/samuelbroscheit/entity_knowledge_in_bert/HEAD/docs/Bert-Entity.png -------------------------------------------------------------------------------- /docs/preprocessing.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/samuelbroscheit/entity_knowledge_in_bert/HEAD/docs/preprocessing.png -------------------------------------------------------------------------------- /downstream_experiments/ensemble_bert_modeling.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/samuelbroscheit/entity_knowledge_in_bert/HEAD/downstream_experiments/ensemble_bert_modeling.py -------------------------------------------------------------------------------- /downstream_experiments/fairseq_patch_01.patch: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/samuelbroscheit/entity_knowledge_in_bert/HEAD/downstream_experiments/fairseq_patch_01.patch -------------------------------------------------------------------------------- /downstream_experiments/fairseq_patch_02.patch: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/samuelbroscheit/entity_knowledge_in_bert/HEAD/downstream_experiments/fairseq_patch_02.patch -------------------------------------------------------------------------------- /downstream_experiments/prepare_fairseq.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/samuelbroscheit/entity_knowledge_in_bert/HEAD/downstream_experiments/prepare_fairseq.sh -------------------------------------------------------------------------------- /downstream_experiments/run_bert_glue_classifier.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/samuelbroscheit/entity_knowledge_in_bert/HEAD/downstream_experiments/run_bert_glue_classifier.py -------------------------------------------------------------------------------- /downstream_experiments/run_bert_glue_classifier_ensemble.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/samuelbroscheit/entity_knowledge_in_bert/HEAD/downstream_experiments/run_bert_glue_classifier_ensemble.py -------------------------------------------------------------------------------- /downstream_experiments/run_bert_squad.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/samuelbroscheit/entity_knowledge_in_bert/HEAD/downstream_experiments/run_bert_squad.py -------------------------------------------------------------------------------- /downstream_experiments/run_bert_squad_ensemble.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/samuelbroscheit/entity_knowledge_in_bert/HEAD/downstream_experiments/run_bert_squad_ensemble.py -------------------------------------------------------------------------------- /downstream_experiments/run_bert_swag_ensemble.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/samuelbroscheit/entity_knowledge_in_bert/HEAD/downstream_experiments/run_bert_swag_ensemble.py -------------------------------------------------------------------------------- /downstream_experiments/run_fairseq_bert_ensemble.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/samuelbroscheit/entity_knowledge_in_bert/HEAD/downstream_experiments/run_fairseq_bert_ensemble.sh -------------------------------------------------------------------------------- /requirements.txt: -------------------------------------------------------------------------------- 1 | flashtext 2 | configargparse 3 | boto3 4 | torch==1.5.0 -------------------------------------------------------------------------------- /setup_paths: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/samuelbroscheit/entity_knowledge_in_bert/HEAD/setup_paths --------------------------------------------------------------------------------