Commits · ee1164d036ce2176ad0fdf55ee84a7a63dcf4324 · nlp / ahisto-modules / Named Entity Experiments · GitLab

Sep 20, 2022
- Use micro-averaged `PER` and `LOC` F1-score as default objective · ee1164d0
  Vít Starý Novotný authored Sep 19, 2022
  
  ee1164d0
Sep 19, 2022
- Use micro-averaging in `AggregateMeanFScoreEvaluator.__call__()` · 51b285b6
  Vít Starý Novotný authored Sep 19, 2022
  
  51b285b6
- Update evaluation of NER models · 152f26f4
  Vít Starý Novotný authored Sep 19, 2022
  
  152f26f4
- Do not expect `category_map` to know all categories in `GROUPS.values()` · d167b7be
  Vít Starý Novotný authored Sep 19, 2022
  
  d167b7be
- Make `AggregateMeanFScoreEvaluator` hashable and totally ordered · b400288e
  Vít Starý Novotný authored Sep 19, 2022
  
  b400288e
- Rename `AggregateMeanFScore` to `AggregateMeanFScoreEvaluator` · b36d1d79
  Vít Starý Novotný authored Sep 19, 2022
  
  b36d1d79
Sep 18, 2022
- Explicitly use sets in `ahisto_named_entity_search.recognition.evaluator` · 6c760a64
  Vít Starý Novotný authored Sep 18, 2022
  
  6c760a64
- Add `*-ORG` and `*-MISC` to `AggregateMeanFScore.GROUPS` · 7e5701eb
  Vít Starý Novotný authored Sep 18, 2022
  
  7e5701eb
Sep 17, 2022
- Add `ahisto_named_entity_search.recognition.evaluator` · 63feb75b
  Vít Starý Novotný authored Sep 16, 2022
  63feb75b
Sep 16, 2022
- Use testing dataset to evaluate NER models · 171526e6
  Vít Starý Novotný authored Sep 16, 2022
  
  171526e6
Sep 15, 2022
- Add TODOs to `03_train_ner_models.ipynb` · 3aa935de
  Vít Starý Novotný authored Sep 15, 2022
  
  3aa935de
- Evaluate baseline model in `03_train_ner_models.ipynb` · 2388bc47
  Vít Starý Novotný authored Sep 15, 2022
  
  2388bc47
- Add `labels` parameter to `NerModel.__init__()` · 2287e9ba
  Vít Starý Novotný authored Sep 15, 2022
  
  2287e9ba
- Add `NerModel.tokenizer` · a15f4ea0
  Vít Starý Novotný authored Sep 15, 2022
  
  a15f4ea0
Sep 13, 2022
- Add evaluation results for all but one model to `03_train_ner_models.ipynb` · 69c56bf1
  Vít Starý Novotný authored Sep 13, 2022
  
  69c56bf1
Sep 12, 2022
- Fix style check · cd193fa8
  Vít Starý Novotný authored Sep 12, 2022
  
  cd193fa8
- Fix type check · f7aef48c
  Vít Starý Novotný authored Sep 12, 2022
  
  f7aef48c
- Add `03_train_ner_models.ipynb` · 7770581d
  Vít Starý Novotný authored Sep 12, 2022
  
  7770581d
- Fix typos in `ahisto_named_entity_search.recognition.model` · abb3e082
  Vít Starý Novotný authored Sep 11, 2022
  
  abb3e082
Sep 11, 2022
- Update `*.ipynb` · 47e65c62
  Vít Starý Novotný authored Sep 11, 2022
  
  47e65c62
- Add `NerModel.test()` · 27e52d2c
  Vít Starý Novotný authored Sep 11, 2022
  
  27e52d2c
- Fix GPU selection script · e8e1d2d4
  Vít Starý Novotný authored Sep 11, 2022
  
  e8e1d2d4
- Make `NerModel.train_and_save()` return model · 98350de5
  Vít Starý Novotný authored Sep 11, 2022
  
  98350de5
- Make `NerModel.__init__()` lazy and add `str()` and `repr()` · 7263450d
  Vít Starý Novotný authored Sep 11, 2022
  
  7263450d
Aug 22, 2022
- Prune old checkpoints · b78567bb
  Vít Starý Novotný authored Aug 22, 2022
  
  b78567bb
Aug 16, 2022
- Fix joblog format in `scripts/03_train_ner_models.sh` · 72d22741
  Vít Starý Novotný authored Aug 16, 2022
  
  72d22741
- Increase delay in `scripts/03_train_ner_models.sh` · 51be26ee
  Vít Starý Novotný authored Aug 16, 2022
  
  51be26ee
- Fix off-by-one error in schedules · 837ffec0
  Vít Starý Novotný authored Aug 16, 2022
  
  837ffec0
- Increase early stopping patience for training NER models · ff8e4281
  Vít Starý Novotný authored Aug 16, 2022
  
  ff8e4281
Aug 15, 2022
- Prevent future objectives from affecting early stopping (cont.) · 832adbaa
  Vít Starý Novotný authored Aug 15, 2022
  
  832adbaa
- Start taking GPUs from the tail of IDs · a3022d22
  Vít Starý Novotný authored Aug 15, 2022
  
  a3022d22
- Remove extra forward slash · eb609384
  Vít Starý Novotný authored Aug 15, 2022
  
  eb609384
- Add GNU parallel joblog to training NER models · 6e472d42
  Vít Starý Novotný authored Aug 15, 2022
  
  6e472d42
- Train using NER tags in BIO format · 1046301d
  Vít Starý Novotný authored Aug 15, 2022
  
  1046301d
- Use 4 jobs to train NER model at most · 16f1e88d
  Vít Starý Novotný authored Aug 15, 2022
  
  16f1e88d
- Prevent future objectives from affecting early stopping · 590372a4
  Vít Starý Novotný authored Aug 15, 2022
  
  590372a4
- Use fine-tuning schedule to train NER models · c4c8ffd3
  Vít Starý Novotný authored Aug 15, 2022
  
  c4c8ffd3
- Evaluate more often with fewer samples · 04b847a2
  Vít Starý Novotný authored Aug 15, 2022
  
  04b847a2
Aug 13, 2022
- Use parallel schedule to train NER models · b3810c32
  Vít Starý Novotný authored Aug 13, 2022
  
  b3810c32
- Separate NerModel training schedules to `recognition.schedule` · 3d99edcc
  Vít Starý Novotný authored Aug 13, 2022
  
  3d99edcc