mbert-tamil-pos-tagging
This model is a fine-tuned version of google-bert/bert-base-multilingual-cased on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 1.2674
- Accuracy: 0.8155
- F1 Macro: 0.6485
- F1 Weighted: 0.8040
- Precision Macro: 0.7457
- Recall Macro: 0.6312
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: 32
- eval_batch_size: 64
- seed: 42
- optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 0.1
- num_epochs: 5
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy | F1 Macro | F1 Weighted | Precision Macro | Recall Macro |
|---|---|---|---|---|---|---|---|---|
| No log | 1.0 | 13 | 3.4280 | 0.5487 | 0.2485 | 0.4814 | 0.3127 | 0.2509 |
| 3.7703 | 2.0 | 26 | 2.0470 | 0.7276 | 0.3839 | 0.6701 | 0.4281 | 0.4114 |
| 3.7703 | 3.0 | 39 | 1.5103 | 0.7838 | 0.5052 | 0.7581 | 0.6704 | 0.5086 |
| 1.7036 | 4.0 | 52 | 1.3280 | 0.8036 | 0.6163 | 0.7903 | 0.7440 | 0.6014 |
| 1.7036 | 5.0 | 65 | 1.2674 | 0.8155 | 0.6485 | 0.8040 | 0.7457 | 0.6312 |
Framework versions
- Transformers 5.0.0
- Pytorch 2.10.0+cu128
- Datasets 5.0.0
- Tokenizers 0.22.2