mbert-targin-final
This model is a fine-tuned version of bert-base-multilingual-cased on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.9847
- Accuracy: 0.7025
- Precision: 0.6490
- Recall: 0.6487
- F1: 0.6489
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: 1e-05
- train_batch_size: 32
- eval_batch_size: 32
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 10
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy | Precision | Recall | F1 |
|---|---|---|---|---|---|---|---|
| No log | 1.0 | 296 | 0.5774 | 0.7091 | 0.6506 | 0.6378 | 0.6426 |
| 0.5912 | 2.0 | 592 | 0.5316 | 0.7376 | 0.6880 | 0.6767 | 0.6814 |
| 0.5912 | 3.0 | 888 | 0.5511 | 0.7253 | 0.6692 | 0.6293 | 0.6378 |
| 0.4844 | 4.0 | 1184 | 0.6262 | 0.6835 | 0.6622 | 0.6884 | 0.6613 |
| 0.4844 | 5.0 | 1480 | 0.6320 | 0.7006 | 0.6574 | 0.6701 | 0.6616 |
| 0.3861 | 6.0 | 1776 | 0.6983 | 0.7148 | 0.6632 | 0.6620 | 0.6626 |
| 0.2773 | 7.0 | 2072 | 0.8109 | 0.7110 | 0.6630 | 0.6689 | 0.6655 |
| 0.2773 | 8.0 | 2368 | 0.8948 | 0.7072 | 0.6525 | 0.6487 | 0.6504 |
| 0.2068 | 9.0 | 2664 | 0.9693 | 0.7072 | 0.6519 | 0.6469 | 0.6492 |
| 0.2068 | 10.0 | 2960 | 0.9847 | 0.7025 | 0.6490 | 0.6487 | 0.6489 |
Framework versions
- Transformers 4.24.0.dev0
- Pytorch 1.11.0+cu102
- Datasets 2.6.1
- Tokenizers 0.13.1