<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. -->
my_ner_model2_1_
This model is a fine-tuned version of dbmdz/bert-base-turkish-cased on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.3810
- Precision: 0.6157
- Recall: 0.75
- F1: 0.6762
- Accuracy: 0.9298
Model description
More information needed
Intended uses & limitations
More information needed
Training and evaluation data
More information needed
Training procedure
Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 2e-05
- train_batch_size: 16
- eval_batch_size: 16
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 20
Training results
| Training Loss | Epoch | Step | Validation Loss | Precision | Recall | F1 | Accuracy |
|---|---|---|---|---|---|---|---|
| No log | 1.0 | 72 | 0.2630 | 0.5114 | 0.6202 | 0.5606 | 0.9178 |
| No log | 2.0 | 144 | 0.2351 | 0.56 | 0.6983 | 0.6216 | 0.9268 |
| No log | 3.0 | 216 | 0.2496 | 0.5645 | 0.7338 | 0.6381 | 0.9243 |
| No log | 4.0 | 288 | 0.2540 | 0.5769 | 0.7440 | 0.6499 | 0.9265 |
| No log | 5.0 | 360 | 0.2559 | 0.6246 | 0.7398 | 0.6773 | 0.9289 |
| No log | 6.0 | 432 | 0.2720 | 0.6137 | 0.7410 | 0.6714 | 0.9297 |
| 0.1489 | 7.0 | 504 | 0.2800 | 0.6305 | 0.7476 | 0.6841 | 0.9312 |
| 0.1489 | 8.0 | 576 | 0.3182 | 0.5866 | 0.7632 | 0.6634 | 0.9251 |
| 0.1489 | 9.0 | 648 | 0.3042 | 0.6127 | 0.75 | 0.6744 | 0.9285 |
| 0.1489 | 10.0 | 720 | 0.3182 | 0.6181 | 0.7518 | 0.6784 | 0.9293 |
| 0.1489 | 11.0 | 792 | 0.3347 | 0.6011 | 0.7506 | 0.6676 | 0.9278 |
| 0.1489 | 12.0 | 864 | 0.3572 | 0.6089 | 0.7596 | 0.6759 | 0.9281 |
| 0.1489 | 13.0 | 936 | 0.3516 | 0.6169 | 0.7548 | 0.6789 | 0.9297 |
| 0.0501 | 14.0 | 1008 | 0.3612 | 0.6175 | 0.7566 | 0.6800 | 0.9294 |
| 0.0501 | 15.0 | 1080 | 0.3683 | 0.6099 | 0.7488 | 0.6722 | 0.9289 |
| 0.0501 | 16.0 | 1152 | 0.3699 | 0.6265 | 0.75 | 0.6827 | 0.9306 |
| 0.0501 | 17.0 | 1224 | 0.3737 | 0.6210 | 0.7542 | 0.6811 | 0.9297 |
| 0.0501 | 18.0 | 1296 | 0.3754 | 0.6211 | 0.7476 | 0.6785 | 0.9291 |
| 0.0501 | 19.0 | 1368 | 0.3823 | 0.6132 | 0.7536 | 0.6762 | 0.9293 |
| 0.0501 | 20.0 | 1440 | 0.3810 | 0.6157 | 0.75 | 0.6762 | 0.9298 |
Framework versions
- Transformers 4.22.1
- Pytorch 1.12.1+cu102
- Datasets 2.7.1
- Tokenizers 0.12.1
