deberta-large-semeval25_EN08_fold2
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: 7.7224
- Precision Samples: 0.1123
- Recall Samples: 0.7856
- F1 Samples: 0.1886
- Precision Macro: 0.3681
- Recall Macro: 0.6639
- F1 Macro: 0.2792
- Precision Micro: 0.1054
- Recall Micro: 0.7394
- F1 Micro: 0.1844
- Precision Weighted: 0.1953
- Recall Weighted: 0.7394
- F1 Weighted: 0.2101
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: 8
- eval_batch_size: 8
- seed: 42
- optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: linear
- num_epochs: 10
Training results
Training Loss | Epoch | Step | Validation Loss | Precision Samples | Recall Samples | F1 Samples | Precision Macro | Recall Macro | F1 Macro | Precision Micro | Recall Micro | F1 Micro | Precision Weighted | Recall Weighted | F1 Weighted |
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
10.1108 | 1.0 | 73 | 9.1532 | 0.1078 | 0.4896 | 0.1658 | 0.8477 | 0.3196 | 0.2145 | 0.1041 | 0.3879 | 0.1641 | 0.6101 | 0.3879 | 0.0866 |
8.9481 | 2.0 | 146 | 8.6922 | 0.1037 | 0.6457 | 0.1672 | 0.7149 | 0.4241 | 0.2235 | 0.0947 | 0.5515 | 0.1616 | 0.4524 | 0.5515 | 0.1227 |
9.1563 | 3.0 | 219 | 8.6496 | 0.0968 | 0.7189 | 0.1614 | 0.5923 | 0.5060 | 0.2388 | 0.0875 | 0.6485 | 0.1542 | 0.3085 | 0.6485 | 0.1447 |
8.7006 | 4.0 | 292 | 8.2522 | 0.1016 | 0.7955 | 0.1617 | 0.5424 | 0.5864 | 0.2606 | 0.0877 | 0.7333 | 0.1567 | 0.2756 | 0.7333 | 0.1672 |
8.1242 | 5.0 | 365 | 7.9321 | 0.1011 | 0.7940 | 0.1721 | 0.4725 | 0.6190 | 0.2653 | 0.0945 | 0.7364 | 0.1675 | 0.2425 | 0.7364 | 0.1754 |
7.4891 | 6.0 | 438 | 8.0728 | 0.1081 | 0.7863 | 0.1824 | 0.4759 | 0.6115 | 0.2650 | 0.0989 | 0.7303 | 0.1743 | 0.2454 | 0.7303 | 0.1816 |
8.3973 | 7.0 | 511 | 7.8203 | 0.1074 | 0.7803 | 0.1817 | 0.3908 | 0.6341 | 0.2637 | 0.1002 | 0.7424 | 0.1765 | 0.1962 | 0.7424 | 0.1906 |
7.0048 | 8.0 | 584 | 7.7429 | 0.1097 | 0.7953 | 0.1849 | 0.3862 | 0.6590 | 0.2731 | 0.1017 | 0.7515 | 0.1791 | 0.2017 | 0.7515 | 0.2014 |
6.3856 | 9.0 | 657 | 7.7281 | 0.1081 | 0.7852 | 0.1823 | 0.3555 | 0.6382 | 0.2597 | 0.1016 | 0.7424 | 0.1788 | 0.1924 | 0.7424 | 0.2033 |
5.8015 | 10.0 | 730 | 7.7224 | 0.1123 | 0.7856 | 0.1886 | 0.3681 | 0.6639 | 0.2792 | 0.1054 | 0.7394 | 0.1844 | 0.1953 | 0.7394 | 0.2101 |
Framework versions
- Transformers 4.46.0
- Pytorch 2.3.1
- Datasets 2.21.0
- Tokenizers 0.20.1
- Downloads last month
- 6
Inference Providers
NEW
This model is not currently available via any of the supported Inference Providers.
Model tree for g-assismoraes/deberta-large-semeval25_EN08_fold2
Base model
microsoft/deberta-v3-large