roberta-base-ag-news-classifier
This model is a fine-tuned version of FacebookAI/roberta-base on fancyzhx/ag_news) dataset. It achieves the following results on the evaluation set:
- Loss: 0.2021
- Accuracy: 0.9541
- F1: 0.9541
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: 32
- 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_ratio: 0.1
- num_epochs: 3
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy | F1 |
|---|---|---|---|---|---|
| 0.2148 | 1.0 | 7500 | 0.1850 | 0.9426 | 0.9425 |
| 0.1646 | 2.0 | 15000 | 0.1778 | 0.9501 | 0.9501 |
| 0.1072 | 3.0 | 22500 | 0.2021 | 0.9541 | 0.9541 |
Framework versions
- Transformers 4.57.1
- Pytorch 2.9.0a0+git1c57644
- Datasets 4.3.0
- Tokenizers 0.22.1