<!-- 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. -->
BerTurk_Electra_15_epoch
This model is a fine-tuned version of dbmdz/electra-base-turkish-cased-discriminator on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.0405
- Accuracy: 0.9931
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: 15
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|---|---|---|---|---|
| No log | 1.0 | 50 | 0.5097 | 0.9722 |
| No log | 2.0 | 100 | 0.1894 | 0.9514 |
| No log | 3.0 | 150 | 0.0747 | 0.9722 |
| No log | 4.0 | 200 | 0.0426 | 0.9792 |
| No log | 5.0 | 250 | 0.0108 | 0.9931 |
| No log | 6.0 | 300 | 0.0321 | 0.9931 |
| No log | 7.0 | 350 | 0.0415 | 0.9861 |
| No log | 8.0 | 400 | 0.0243 | 0.9931 |
| No log | 9.0 | 450 | 0.0319 | 0.9931 |
| 0.0967 | 10.0 | 500 | 0.0351 | 0.9931 |
| 0.0967 | 11.0 | 550 | 0.0369 | 0.9931 |
| 0.0967 | 12.0 | 600 | 0.0387 | 0.9931 |
| 0.0967 | 13.0 | 650 | 0.0397 | 0.9931 |
| 0.0967 | 14.0 | 700 | 0.0404 | 0.9931 |
| 0.0967 | 15.0 | 750 | 0.0405 | 0.9931 |
Framework versions
- Transformers 4.25.1
- Pytorch 1.13.0+cu116
- Datasets 2.8.0
- Tokenizers 0.13.2
