metadata
library_name: transformers
license: mit
base_model: dbmdz/bert-base-turkish-cased
tags:
- generated_from_trainer
metrics:
- f1
model-index:
- name: turkish-medical-question-answering
results: []
turkish-medical-question-answering
This model is a fine-tuned version of dbmdz/bert-base-turkish-cased on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 1.2864
- Exact Match: 55.1899
- F1: 75.1246
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: 1e-05
- train_batch_size: 16
- eval_batch_size: 64
- seed: 42
- optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: cosine
- lr_scheduler_warmup_steps: 1000
- num_epochs: 10
Training results
| Training Loss | Epoch | Step | Validation Loss | Exact Match | F1 |
|---|---|---|---|---|---|
| 5.9451 | 0.1166 | 50 | 5.9731 | 0.0 | 5.7871 |
| 5.8483 | 0.2331 | 100 | 5.8315 | 0.0 | 5.9124 |
| 5.6758 | 0.3497 | 150 | 5.6207 | 0.2519 | 6.2966 |
| 5.4079 | 0.4662 | 200 | 5.3396 | 0.2584 | 6.1159 |
| 5.1136 | 0.5828 | 250 | 5.0135 | 0.2755 | 7.7324 |
| 4.7974 | 0.6993 | 300 | 4.6818 | 1.2531 | 9.4001 |
| 4.4412 | 0.8159 | 350 | 4.3335 | 1.5656 | 10.9210 |
| 4.1566 | 0.9324 | 400 | 3.8584 | 7.0281 | 22.6904 |
| 3.599 | 1.0490 | 450 | 2.8534 | 23.4628 | 48.7349 |
| 3.0347 | 1.1655 | 500 | 2.4007 | 28.9362 | 53.4749 |
| 2.7626 | 1.2821 | 550 | 2.2064 | 33.4711 | 57.0427 |
| 2.5741 | 1.3986 | 600 | 2.0689 | 37.3297 | 60.0019 |
| 2.3917 | 1.5152 | 650 | 1.9761 | 38.6968 | 61.5842 |
| 2.3607 | 1.6317 | 700 | 1.9137 | 41.5344 | 63.5713 |
| 2.2972 | 1.7483 | 750 | 1.8294 | 44.0210 | 65.5393 |
| 2.1214 | 1.8648 | 800 | 1.7721 | 44.7028 | 65.9376 |
| 2.1775 | 1.9814 | 850 | 1.7058 | 46.1538 | 66.8453 |
| 1.9282 | 2.0979 | 900 | 1.6579 | 46.7784 | 67.7813 |
| 1.9428 | 2.2145 | 950 | 1.6366 | 46.8912 | 68.5467 |
| 1.8639 | 2.3310 | 1000 | 1.5751 | 49.4819 | 70.2320 |
| 1.8969 | 2.4476 | 1050 | 1.5519 | 48.9610 | 70.3424 |
| 1.7348 | 2.5641 | 1100 | 1.5173 | 50.4505 | 71.1044 |
| 1.7847 | 2.6807 | 1150 | 1.4999 | 50.5762 | 71.6186 |
| 1.7822 | 2.7972 | 1200 | 1.4566 | 53.0691 | 72.0005 |
| 1.7989 | 2.9138 | 1250 | 1.4300 | 51.6005 | 72.0121 |
| 1.7683 | 3.0303 | 1300 | 1.4319 | 52.0305 | 72.2366 |
| 1.5444 | 3.1469 | 1350 | 1.4277 | 51.7903 | 72.0603 |
| 1.5121 | 3.2634 | 1400 | 1.3861 | 53.6122 | 73.3486 |
| 1.6294 | 3.3800 | 1450 | 1.3830 | 52.6718 | 73.2456 |
| 1.514 | 3.4965 | 1500 | 1.3456 | 53.7389 | 73.4757 |
| 1.3778 | 3.6131 | 1550 | 1.3644 | 53.2319 | 73.5271 |
| 1.4502 | 3.7296 | 1600 | 1.3491 | 53.6030 | 73.8642 |
| 1.5388 | 3.8462 | 1650 | 1.3611 | 53.0380 | 73.1390 |
| 1.5244 | 3.9627 | 1700 | 1.3143 | 53.3587 | 74.0002 |
| 1.3127 | 4.0793 | 1750 | 1.3191 | 54.4767 | 74.6247 |
| 1.3819 | 4.1958 | 1800 | 1.2864 | 55.1899 | 75.1246 |
| 1.307 | 4.3124 | 1850 | 1.3762 | 54.1401 | 74.4158 |
| 1.2792 | 4.4289 | 1900 | 1.3156 | 53.4943 | 75.0122 |
| 1.289 | 4.5455 | 1950 | 1.2809 | 55.0063 | 74.6137 |
Framework versions
- Transformers 4.48.0.dev0
- Pytorch 2.4.1+cu121
- Datasets 3.1.0
- Tokenizers 0.21.0