xlm-roberta-base
This model was trained from scratch on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.1437
- F1: 0.6011
- Roc Auc: 0.7429
- Accuracy: 0.4893
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: 8
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 3
Training results
| Training Loss | Epoch | Step | Validation Loss | F1 | Roc Auc | Accuracy |
|---|---|---|---|---|---|---|
| 0.1911 | 1.0 | 1336 | 0.1682 | 0.5027 | 0.6816 | 0.3685 |
| 0.1506 | 2.0 | 2672 | 0.1485 | 0.5670 | 0.7204 | 0.4489 |
| 0.1379 | 3.0 | 4008 | 0.1437 | 0.6011 | 0.7429 | 0.4893 |
Framework versions
- Transformers 4.25.1
- Pytorch 1.13.0+cu116
- Datasets 2.8.0
- Tokenizers 0.13.2