german-sentiment-bert-finetuned-augmentation-deu-MICRO
This model is a fine-tuned version of sercetexam9/german-sentiment-bert-finetuned-augmentation on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.3195
- F1: 0.7616
- Roc Auc: 0.8429
- Accuracy: 0.6059
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: cosine
- lr_scheduler_warmup_steps: 100
- num_epochs: 20
Training results
| Training Loss | Epoch | Step | Validation Loss | F1 | Roc Auc | Accuracy |
|---|---|---|---|---|---|---|
| 0.1945 | 1.0 | 846 | 0.2829 | 0.7454 | 0.8366 | 0.5791 |
| 0.1007 | 2.0 | 1692 | 0.2853 | 0.7611 | 0.8356 | 0.6105 |
| 0.06 | 3.0 | 2538 | 0.3195 | 0.7616 | 0.8429 | 0.6059 |
| 0.0362 | 4.0 | 3384 | 0.3590 | 0.7416 | 0.8177 | 0.5874 |
| 0.0337 | 5.0 | 4230 | 0.3966 | 0.7565 | 0.8351 | 0.5957 |
| 0.0209 | 6.0 | 5076 | 0.4195 | 0.7545 | 0.8441 | 0.5856 |
| 0.0132 | 7.0 | 5922 | 0.4535 | 0.7589 | 0.8425 | 0.5883 |
Framework versions
- Transformers 4.45.1
- Pytorch 2.4.0
- Datasets 3.0.1
- Tokenizers 0.20.0