berturk-uncased-keyword-discriminator
This model is a fine-tuned version of dbmdz/bert-base-turkish-uncased on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.3989
- Precision: 0.6234
- Recall: 0.6508
- Accuracy: 0.9145
- F1: 0.6368
- Ent/precision: 0.6435
- Ent/accuracy: 0.7120
- Ent/f1: 0.6761
- Con/precision: 0.5834
- Con/accuracy: 0.5475
- Con/f1: 0.5649
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 | Ent/precision | Ent/accuracy | Ent/f1 | Con/precision | Con/accuracy | Con/f1 |
---|---|---|---|---|---|---|---|---|---|---|---|---|---|
0.2005 | 1.0 | 1875 | 0.2104 | 0.5981 | 0.5978 | 0.9148 | 0.5979 | 0.6280 | 0.6665 | 0.6467 | 0.5383 | 0.4820 | 0.5086 |
0.1468 | 2.0 | 3750 | 0.2094 | 0.5996 | 0.6568 | 0.9164 | 0.6269 | 0.6285 | 0.7049 | 0.6645 | 0.5477 | 0.5757 | 0.5614 |
0.1124 | 3.0 | 5625 | 0.2372 | 0.6106 | 0.6679 | 0.9154 | 0.6380 | 0.6285 | 0.7270 | 0.6741 | 0.5753 | 0.5684 | 0.5718 |
0.0861 | 4.0 | 7500 | 0.2736 | 0.6133 | 0.6707 | 0.9145 | 0.6407 | 0.6281 | 0.7359 | 0.6777 | 0.5830 | 0.5606 | 0.5716 |
0.0644 | 5.0 | 9375 | 0.3081 | 0.6115 | 0.6683 | 0.9145 | 0.6386 | 0.6291 | 0.7293 | 0.6755 | 0.5764 | 0.5657 | 0.5710 |
0.0498 | 6.0 | 11250 | 0.3449 | 0.6245 | 0.6466 | 0.9149 | 0.6353 | 0.6380 | 0.7097 | 0.6720 | 0.5965 | 0.5401 | 0.5669 |
0.0401 | 7.0 | 13125 | 0.3838 | 0.6223 | 0.6545 | 0.9140 | 0.6380 | 0.6449 | 0.7100 | 0.6759 | 0.5790 | 0.5610 | 0.5699 |
0.0329 | 8.0 | 15000 | 0.3989 | 0.6234 | 0.6508 | 0.9145 | 0.6368 | 0.6435 | 0.7120 | 0.6761 | 0.5834 | 0.5475 | 0.5649 |
Framework versions
- Transformers 4.19.2
- Pytorch 1.11.0+cu113
- Datasets 2.2.2
- Tokenizers 0.12.1
- Downloads last month
- 30
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social
visibility and check back later, or deploy to Inference Endpoints (dedicated)
instead.