convberturk-keyword-extractor
This model is a fine-tuned version of dbmdz/convbert-base-turkish-cased on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.4098
- Precision: 0.6742
- Recall: 0.7035
- Accuracy: 0.9175
- F1: 0.6886
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: 8
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Precision | Recall | Accuracy | F1 |
|---|---|---|---|---|---|---|---|
| 0.174 | 1.0 | 1875 | 0.1920 | 0.6546 | 0.6869 | 0.9184 | 0.6704 |
| 0.1253 | 2.0 | 3750 | 0.2030 | 0.6527 | 0.7317 | 0.9179 | 0.6900 |
| 0.091 | 3.0 | 5625 | 0.2517 | 0.6499 | 0.7473 | 0.9163 | 0.6952 |
| 0.0684 | 4.0 | 7500 | 0.2828 | 0.6633 | 0.7270 | 0.9167 | 0.6937 |
| 0.0536 | 5.0 | 9375 | 0.3307 | 0.6706 | 0.7194 | 0.9180 | 0.6942 |
| 0.0384 | 6.0 | 11250 | 0.3669 | 0.6655 | 0.7161 | 0.9157 | 0.6898 |
| 0.0316 | 7.0 | 13125 | 0.3870 | 0.6792 | 0.7002 | 0.9176 | 0.6895 |
| 0.0261 | 8.0 | 15000 | 0.4098 | 0.6742 | 0.7035 | 0.9175 | 0.6886 |
Framework versions
- Transformers 4.19.2
- Pytorch 1.11.0+cu113
- Datasets 2.2.2
- Tokenizers 0.12.1