Instructions to use kali1/KLUE-custom-ner with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use kali1/KLUE-custom-ner with Transformers:
# pip install -U transformers accelerate # Load model directly from transformers import AutoTokenizer, BertForEntity tokenizer = AutoTokenizer.from_pretrained("kali1/KLUE-custom-ner") model = BertForEntity.from_pretrained("kali1/KLUE-custom-ner", device_map="auto") - Notebooks
- Google Colab
- Kaggle
Download eval.log from kali1/KLUE-custom-ner: direct link, hf CLI and curl.
- Browser
- Download file 6.27 kB
-
https://huggingface.co/kali1/KLUE-custom-ner/resolve/main/eval.log
- Command line
-
hf download hf://kali1/KLUE-custom-ner/eval.log
-
curl -L -o eval.log https://huggingface.co/kali1/KLUE-custom-ner/resolve/main/eval.log
6.27 kB
| ['run_entity.py', '--do_eval', '--eval_test', '--learning_rate=1e-5', '--task_learning_rate=5e-4', '--train_batch_size=32', '--context_window', '0', '--task', 'ko', '--data_dir', 'korean_data/json/', '--model', 'monologg/kobert', '--output_dir', 'sh_aug+klue_dir'] | |
| Namespace(bert_model_dir=None, bertadam=False, context_window=0, data_dir='korean_data/json/', dev_data='korean_data/json/dev.json', dev_pred_filename='ent_pred_dev.json', do_eval=True, do_train=False, eval_batch_size=32, eval_per_epoch=1, eval_test=True, learning_rate=1e-05, max_span_length=8, model='monologg/kobert', num_epoch=20, output_dir='sh_aug+klue_dir', print_loss_step=100, seed=0, task='ko', task_learning_rate=0.0005, test_data='korean_data/json/test.json', test_pred_filename='ent_pred_test.json', train_batch_size=32, train_data='korean_data/json/train.json', train_shuffle=False, use_albert=False, warmup_proportion=0.1) | |
| Moving to CUDA... | |
| # GPUs = 1 | |
| # Overlap: 0 | |
| Extracted 5000 samples from 5000 documents, with 14178 NER labels, 31.704 avg input length, 93 max length | |
| Max Length: 93, max NER: 16 | |
| Loading BERT model from sh_aug+klue_dir/ | |
| Moving to CUDA... | |
| # GPUs = 1 | |
| Loading BERT model from KLUE_dir/ | |
| Moving to CUDA... | |
| # GPUs = 1 | |
| # Overlap: 0 | |
| Extracted 1 samples from 1 documents, with 0 NER labels, 15.000 avg input length, 15 max length | |
| Max Length: 15, max NER: 0 | |
| Evaluating... | |
| Accuracy: 0.978261 | |
| Cor: 0, Pred TOT: 2, Gold TOT: 0 | |
| P: 0.00000, R: 0.00000, F1: 0.00000 | |
| Used time: 1.660250 | |
| Evaluating... | |
| Accuracy: 0.989130 | |
| Cor: 0, Pred TOT: 1, Gold TOT: 0 | |
| P: 0.00000, R: 0.00000, F1: 0.00000 | |
| Used time: 0.009506 | |
| Total pred entities: 2 | |
| Output predictions to sh_aug+klue_dir\ent_pred_test.json.. | |
| Total pred entities: 1 | |
| Output predictions to sh_aug+klue_dir\ent_pred_test.json.. | |
| # Overlap: 0 | |
| Extracted 1 samples from 1 documents, with 0 NER labels, 14.000 avg input length, 14 max length | |
| Max Length: 14, max NER: 0 | |
| Evaluating... | |
| Accuracy: 0.952381 | |
| Cor: 0, Pred TOT: 4, Gold TOT: 0 | |
| P: 0.00000, R: 0.00000, F1: 0.00000 | |
| Used time: 0.145730 | |
| Evaluating... | |
| Accuracy: 0.976190 | |
| Cor: 0, Pred TOT: 2, Gold TOT: 0 | |
| P: 0.00000, R: 0.00000, F1: 0.00000 | |
| Used time: 0.203033 | |
| Total pred entities: 4 | |
| Output predictions to sh_aug+klue_dir\ent_pred_test.json.. | |
| Total pred entities: 2 | |
| Output predictions to sh_aug+klue_dir\ent_pred_test.json.. | |
| # Overlap: 0 | |
| Extracted 1 samples from 1 documents, with 0 NER labels, 11.000 avg input length, 11 max length | |
| Max Length: 11, max NER: 0 | |
| Evaluating... | |
| Accuracy: 0.950000 | |
| Cor: 0, Pred TOT: 3, Gold TOT: 0 | |
| P: 0.00000, R: 0.00000, F1: 0.00000 | |
| Used time: 0.149735 | |
| Evaluating... | |
| Accuracy: 0.950000 | |
| Cor: 0, Pred TOT: 3, Gold TOT: 0 | |
| P: 0.00000, R: 0.00000, F1: 0.00000 | |
| Used time: 0.123036 | |
| Total pred entities: 3 | |
| Output predictions to sh_aug+klue_dir\ent_pred_test.json.. | |
| Total pred entities: 3 | |
| Output predictions to sh_aug+klue_dir\ent_pred_test.json.. | |
| # Overlap: 0 | |
| Extracted 1 samples from 1 documents, with 0 NER labels, 10.000 avg input length, 10 max length | |
| Max Length: 10, max NER: 0 | |
| Evaluating... | |
| Accuracy: 0.980769 | |
| Cor: 0, Pred TOT: 1, Gold TOT: 0 | |
| P: 0.00000, R: 0.00000, F1: 0.00000 | |
| Used time: 0.070563 | |
| Evaluating... | |
| Accuracy: 0.980769 | |
| Cor: 0, Pred TOT: 1, Gold TOT: 0 | |
| P: 0.00000, R: 0.00000, F1: 0.00000 | |
| Used time: 0.007507 | |
| Total pred entities: 1 | |
| Output predictions to sh_aug+klue_dir\ent_pred_test.json.. | |
| Total pred entities: 1 | |
| Output predictions to sh_aug+klue_dir\ent_pred_test.json.. | |
| # Overlap: 0 | |
| Extracted 1 samples from 1 documents, with 0 NER labels, 13.000 avg input length, 13 max length | |
| Max Length: 13, max NER: 0 | |
| Evaluating... | |
| Accuracy: 0.986842 | |
| Cor: 0, Pred TOT: 1, Gold TOT: 0 | |
| P: 0.00000, R: 0.00000, F1: 0.00000 | |
| Used time: 0.011518 | |
| Evaluating... | |
| Accuracy: 0.986842 | |
| Cor: 0, Pred TOT: 1, Gold TOT: 0 | |
| P: 0.00000, R: 0.00000, F1: 0.00000 | |
| Used time: 0.010508 | |
| Total pred entities: 1 | |
| Output predictions to sh_aug+klue_dir\ent_pred_test.json.. | |
| Total pred entities: 1 | |
| Output predictions to sh_aug+klue_dir\ent_pred_test.json.. | |
| # Overlap: 0 | |
| Extracted 1 samples from 1 documents, with 0 NER labels, 12.000 avg input length, 12 max length | |
| Max Length: 12, max NER: 0 | |
| Evaluating... | |
| Accuracy: 0.955882 | |
| Cor: 0, Pred TOT: 3, Gold TOT: 0 | |
| P: 0.00000, R: 0.00000, F1: 0.00000 | |
| Used time: 0.147390 | |
| Evaluating... | |
| Accuracy: 0.955882 | |
| Cor: 0, Pred TOT: 3, Gold TOT: 0 | |
| P: 0.00000, R: 0.00000, F1: 0.00000 | |
| Used time: 0.125365 | |
| Total pred entities: 3 | |
| Output predictions to sh_aug+klue_dir\ent_pred_test.json.. | |
| Total pred entities: 3 | |
| Output predictions to sh_aug+klue_dir\ent_pred_test.json.. | |
| # Overlap: 0 | |
| Extracted 1 samples from 1 documents, with 0 NER labels, 12.000 avg input length, 12 max length | |
| Max Length: 12, max NER: 0 | |
| Evaluating... | |
| Accuracy: 0.985294 | |
| Cor: 0, Pred TOT: 1, Gold TOT: 0 | |
| P: 0.00000, R: 0.00000, F1: 0.00000 | |
| Used time: 0.149942 | |
| Evaluating... | |
| Accuracy: 0.985294 | |
| Cor: 0, Pred TOT: 1, Gold TOT: 0 | |
| P: 0.00000, R: 0.00000, F1: 0.00000 | |
| Used time: 0.136553 | |
| Total pred entities: 1 | |
| Output predictions to sh_aug+klue_dir\ent_pred_test.json.. | |
| Total pred entities: 1 | |
| Output predictions to sh_aug+klue_dir\ent_pred_test.json.. | |
| # Overlap: 0 | |
| Extracted 1 samples from 1 documents, with 0 NER labels, 11.000 avg input length, 11 max length | |
| Max Length: 11, max NER: 0 | |
| Evaluating... | |
| Accuracy: 0.966667 | |
| Cor: 0, Pred TOT: 2, Gold TOT: 0 | |
| P: 0.00000, R: 0.00000, F1: 0.00000 | |
| Used time: 0.144237 | |
| Evaluating... | |
| Accuracy: 0.983333 | |
| Cor: 0, Pred TOT: 1, Gold TOT: 0 | |
| P: 0.00000, R: 0.00000, F1: 0.00000 | |
| Used time: 0.133758 | |
| Total pred entities: 2 | |
| Output predictions to sh_aug+klue_dir\ent_pred_test.json.. | |
| Total pred entities: 1 | |
| Output predictions to sh_aug+klue_dir\ent_pred_test.json.. | |
| # Overlap: 0 | |
| Extracted 1 samples from 1 documents, with 0 NER labels, 12.000 avg input length, 12 max length | |
| Max Length: 12, max NER: 0 | |
| Evaluating... | |
| Accuracy: 0.955882 | |
| Cor: 0, Pred TOT: 3, Gold TOT: 0 | |
| P: 0.00000, R: 0.00000, F1: 0.00000 | |
| Used time: 0.080669 | |
| Evaluating... | |
| Accuracy: 0.970588 | |
| Cor: 0, Pred TOT: 2, Gold TOT: 0 | |
| P: 0.00000, R: 0.00000, F1: 0.00000 | |
| Used time: 0.015511 | |
| Total pred entities: 3 | |
| Output predictions to sh_aug+klue_dir\ent_pred_test.json.. | |
| Total pred entities: 2 | |
| Output predictions to sh_aug+klue_dir\ent_pred_test.json.. | |