<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. -->
roberta-large-finetuned-ADEs_model_2
This model is a fine-tuned version of roberta-large on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.2580
- Precision: 0.5407
- Recall: 0.6311
- F1: 0.5824
- Accuracy: 0.8897
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: 5e-07
- 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: 5
Training results
| Training Loss | Epoch | Step | Validation Loss | Precision | Recall | F1 | Accuracy |
|---|---|---|---|---|---|---|---|
| 0.7461 | 1.0 | 640 | 0.3393 | 0.4247 | 0.5095 | 0.4633 | 0.8648 |
| 0.3632 | 2.0 | 1280 | 0.2822 | 0.4934 | 0.6035 | 0.5429 | 0.8819 |
| 0.3102 | 3.0 | 1920 | 0.2663 | 0.5218 | 0.6112 | 0.5630 | 0.8879 |
| 0.2806 | 4.0 | 2560 | 0.2604 | 0.5337 | 0.6311 | 0.5783 | 0.8890 |
| 0.2772 | 5.0 | 3200 | 0.2580 | 0.5407 | 0.6311 | 0.5824 | 0.8897 |
Framework versions
- Transformers 4.18.0
- Pytorch 1.11.0+cu113
- Datasets 2.1.0
- Tokenizers 0.12.1
