deberta-v3-large-finetuned-augmentation-LUNAR
This model is a fine-tuned version of microsoft/deberta-v3-large on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.5590
- F1: 0.8019
- Roc Auc: 0.8448
- Accuracy: 0.5716
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: 16
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_steps: 100
- num_epochs: 20
Training results
| Training Loss | Epoch | Step | Validation Loss | F1 | Roc Auc | Accuracy |
|---|---|---|---|---|---|---|
| 0.488 | 1.0 | 178 | 0.4606 | 0.5349 | 0.6864 | 0.3497 |
| 0.3597 | 2.0 | 356 | 0.3842 | 0.6421 | 0.7471 | 0.4242 |
| 0.2416 | 3.0 | 534 | 0.3560 | 0.7257 | 0.7918 | 0.4930 |
| 0.1479 | 4.0 | 712 | 0.3861 | 0.7658 | 0.8271 | 0.4874 |
| 0.083 | 5.0 | 890 | 0.3829 | 0.7924 | 0.8342 | 0.5506 |
| 0.0488 | 6.0 | 1068 | 0.4568 | 0.7657 | 0.8193 | 0.5225 |
| 0.0339 | 7.0 | 1246 | 0.4678 | 0.7873 | 0.8391 | 0.5534 |
| 0.0164 | 8.0 | 1424 | 0.5445 | 0.7860 | 0.8439 | 0.5421 |
| 0.0108 | 9.0 | 1602 | 0.5306 | 0.7965 | 0.8422 | 0.5744 |
| 0.0046 | 10.0 | 1780 | 0.5590 | 0.8019 | 0.8448 | 0.5716 |
| 0.0032 | 11.0 | 1958 | 0.6109 | 0.7929 | 0.8358 | 0.5562 |
| 0.0072 | 12.0 | 2136 | 0.5936 | 0.7947 | 0.8391 | 0.5562 |
| 0.0014 | 13.0 | 2314 | 0.5925 | 0.7978 | 0.8423 | 0.5688 |
| 0.0011 | 14.0 | 2492 | 0.5986 | 0.7982 | 0.8449 | 0.5618 |
Framework versions
- Transformers 4.45.1
- Pytorch 2.4.0
- Datasets 3.0.1
- Tokenizers 0.20.0