newsdata-bert
This model is a fine-tuned version of bert-base-cased on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.4835
- Accuracy: 0.8617
- Precision: 0.8494
- F1: 0.8533
- Recall: 0.8617
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: 4
- eval_batch_size: 2
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 16
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 1
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy | Precision | F1 | Recall |
|---|---|---|---|---|---|---|---|
| 1.2095 | 0.1024 | 1000 | 1.0182 | 0.7335 | 0.6811 | 0.6915 | 0.7335 |
| 0.8995 | 0.2048 | 2000 | 0.8102 | 0.7798 | 0.7622 | 0.7586 | 0.7798 |
| 0.7554 | 0.3071 | 3000 | 0.6720 | 0.8165 | 0.7938 | 0.8023 | 0.8165 |
| 0.6805 | 0.4095 | 4000 | 0.6185 | 0.828 | 0.8107 | 0.8157 | 0.828 |
| 0.6192 | 0.5119 | 5000 | 0.5865 | 0.8322 | 0.8233 | 0.8226 | 0.8322 |
| 0.5963 | 0.6143 | 6000 | 0.5462 | 0.8475 | 0.8333 | 0.8356 | 0.8475 |
| 0.5466 | 0.7166 | 7000 | 0.5384 | 0.849 | 0.8386 | 0.8398 | 0.849 |
| 0.5447 | 0.8190 | 8000 | 0.4923 | 0.8582 | 0.8440 | 0.8491 | 0.8582 |
| 0.5288 | 0.9214 | 9000 | 0.4835 | 0.8617 | 0.8494 | 0.8533 | 0.8617 |
Framework versions
- Transformers 4.41.2
- Pytorch 2.3.1+cu121
- Datasets 2.20.0
- Tokenizers 0.19.1