Commits · f58bfc3c953b9398bf1df023cd6e7930b2350dd1 · nlp / ahisto-modules / Named Entity Experiments · GitLab

Jan 12, 2023
- Finish evaluation of sixth-generation of models · f58bfc3c
  Vít Starý Novotný authored Jan 12, 2023
  
  f58bfc3c
- Add `search.evaluation.NEREvaluationResult` · ce15d30d
  Vít Starý Novotný authored Jan 11, 2023
  
  ce15d30d
Jan 05, 2023
- Use seventh-generation models to tag MLM datasets · 4b44e99d
  Vít Starý Novotný authored Jan 05, 2023
  
  4b44e99d
Jan 04, 2023
- Update to Python 3.8 · 3698733e
  Vít Starý Novotný authored Jan 04, 2023
  
  3698733e
- Extend quantitative evaluation with confusion matrices · ceb57116
  Vít Starý Novotný authored Jan 04, 2023
  
  ceb57116
- Extend qualitative evaluation to Latin and German example sentences · 3dc65ece
  Vít Starý Novotný authored Jan 04, 2023
  
  3dc65ece
Jan 03, 2023
- Evaluate seventh-generation models · 720fce41
  Vít Starý Novotný authored Jan 03, 2023
  
  720fce41
Dec 30, 2022
- Make checkpoint directory name unique to a Git commit · c7568c43
  Vít Starý Novotný authored Dec 30, 2022
  
  c7568c43
- Increase `warmup_ratio` to 0.1 · e87e66ec
  Vít Starý Novotný authored Dec 30, 2022
  
  e87e66ec
- Add `warmup_ratio` to `config.ini` · 90fee260
  Vít Starý Novotný authored Dec 30, 2022
  
  90fee260
- Set `max_steps` and `number_of_training_epochs` to approximately 200 epochs · 661c6101
  Vít Starý Novotný authored Dec 30, 2022
  
  661c6101
- Work around https://github.com/gaussalgo/adaptor/issues/15 · 3b533d46
  Vít Starý Novotný authored Dec 30, 2022
  
  3b533d46
Dec 29, 2022
- Train seventh-generation models · 037f8531
  Vít Starý Novotný authored Dec 07, 2022
  
  037f8531
Dec 28, 2022
- Evaluate sixth-generation models · d1b65369
  Vít Starý Novotný authored Dec 28, 2022
  
  d1b65369
Dec 27, 2022
- Reduce the number of saved checkpoints · 12ba560d
  Vít Starý Novotný authored Dec 27, 2022
  
  12ba560d
Dec 18, 2022
- Pin `transformers` package · 18c590c4
  Vít Starý Novotný authored Dec 18, 2022
  
  18c590c4
- Update `04_extract_dataset_statistics.ipynb` · 3297b04e
  Vít Starý Novotný authored Dec 18, 2022
  
  3297b04e
- Train sixth-generation models · 58ee3ea0
  Vít Starý Novotný authored Dec 18, 2022
  
  58ee3ea0
- Evaluate fifth-generation models · 8c95e445
  Vít Starý Novotný authored Dec 18, 2022
  
  8c95e445
Dec 16, 2022
- Increase delay between model training · 76e2de6e
  Vít Starý Novotný authored Dec 16, 2022
  
  76e2de6e
- Revert "Load best model at the end of training" · f7c4d3f6
  Vít Starý Novotný authored Dec 16, 2022
```
This reverts commit 44a6597e.

I received the following error when training:
https://discuss.huggingface.co/t/distributed-training-keyerror-eval-f1-in-questionansweringtrainer-taken-from-trainer-qa-py-in-examples/26308
```
  f7c4d3f6
- Train fifth-generation models · a944e451
  Vít Starý Novotný authored Dec 16, 2022
  
  a944e451
- Use fourth-generation models to tag MLM datasets · 027c2b5a
  Vít Starý Novotný authored Dec 16, 2022
  
  027c2b5a
Dec 14, 2022
- Do extra legwork to prevent contamination with test dataset · 4803f5c2
  Vít Starý Novotný authored Dec 14, 2022
  
  4803f5c2
Dec 13, 2022
- Complete MLM datasets using fourth-generation models · 30a72ba2
  Vít Starý Novotný authored Dec 13, 2022
  
  30a72ba2
- Add `TaggedSentence.{__eq__,__hash__}` · f22a29fc
  Vít Starý Novotný authored Dec 13, 2022
  
  f22a29fc
Dec 12, 2022
- Extract dataset statistics · 93ec3431
  Vít Starý Novotný authored Dec 12, 2022
  
  93ec3431
Dec 08, 2022
- Fix a typo · 794b0d1f
  Vít Starý Novotný authored Dec 08, 2022
  
  794b0d1f
- Add learning rate (5e-5, the default value) to config · 58bebe8a
  Vít Starý Novotný authored Dec 08, 2022
  
  58bebe8a
- Fix division by zero in NER evaluation · 176559bb
  Vít Starý Novotný authored Dec 08, 2022
  
  176559bb
- Load best model at the end of training · 44a6597e
  Vít Starý Novotný authored Dec 08, 2022
  
  44a6597e
Dec 07, 2022
- Evaluate fourth-generation models · 02ddb2aa
  Vít Starý Novotný authored Dec 07, 2022
  
  02ddb2aa
Nov 28, 2022
- Increase number of jobs in `scripts/03a_train_ner_models.sh` · f92c6984
  Vít Starý Novotný authored Nov 28, 2022
  
  f92c6984
- Decrease number of jobs in `scripts/03a_train_ner_models.sh` · a4c3b3ab
  Vít Starý Novotný authored Nov 28, 2022
  
  a4c3b3ab
- Increase number of jobs in `scripts/03a_train_ner_models.sh` · afc97be0
  Vít Starý Novotný authored Nov 28, 2022
  
  afc97be0
Nov 27, 2022
- Train fourth generation of NER models · d94fd7bd
  Vít Starý Novotný authored Nov 27, 2022
  
  d94fd7bd
- Complete the train and validation NER datasets · 1e9b05bd
  Vít Starý Novotný authored Nov 26, 2022
  
  1e9b05bd
Nov 25, 2022
- Add `TaggedSentence.density` · 03a0d1a3
  Vít Starý Novotný authored Nov 25, 2022
  
  03a0d1a3
Nov 23, 2022
- Rerun quantitative evaluation · 718820e4
  Vít Starý Novotný authored Nov 23, 2022
  
  718820e4
- Fix type error · e7e7ad3a
  Vít Starý Novotný authored Nov 23, 2022
  
  e7e7ad3a