deberta-v3-large-survey-new_fact_related_passage-rater-all-gpt4
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: 0.4201
- Krippendorff: 0.8827
- Spearman: 0.8845
- Absolute Agreement: 0.8537
- Agreement Within One: 0.9608
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: 6e-06
- 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: linear
- lr_scheduler_warmup_steps: 1000
- num_epochs: 20
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Krippendorff | Spearman | Absolute Agreement | Agreement Within One |
|---|---|---|---|---|---|---|---|
| No log | 1.0 | 55 | 1.9465 | 0.0364 | 0.2240 | 0.0556 | 0.8611 |
| No log | 2.0 | 110 | 1.9474 | 0.0392 | 0.1845 | 0.0556 | 0.7917 |
| No log | 3.0 | 165 | 1.9468 | -0.0269 | nan | 0.0278 | 0.7083 |
| No log | 4.0 | 220 | 1.9454 | -0.0269 | nan | 0.0278 | 0.7083 |
| No log | 5.0 | 275 | 1.9541 | -0.0269 | nan | 0.0278 | 0.7083 |
| No log | 6.0 | 330 | 1.9415 | 0.1181 | 0.2629 | 0.0972 | 0.8194 |
| No log | 7.0 | 385 | 1.9009 | 0.2522 | 0.2306 | 0.2361 | 0.7222 |
| No log | 8.0 | 440 | 1.7945 | 0.1718 | 0.2584 | 0.3194 | 0.6667 |
| No log | 9.0 | 495 | 1.7244 | 0.1199 | 0.2583 | 0.3611 | 0.6528 |
| 1.4338 | 10.0 | 550 | 1.6929 | 0.2689 | 0.2958 | 0.3194 | 0.7083 |
| 1.4338 | 11.0 | 605 | 1.6008 | 0.5568 | 0.5440 | 0.4306 | 0.8333 |
| 1.4338 | 12.0 | 660 | 1.6055 | 0.2689 | 0.3858 | 0.4028 | 0.6667 |
| 1.4338 | 13.0 | 715 | 1.6136 | 0.1576 | 0.3055 | 0.4028 | 0.6528 |
| 1.4338 | 14.0 | 770 | 1.6585 | 0.5199 | 0.5558 | 0.4583 | 0.8194 |
| 1.4338 | 15.0 | 825 | 1.4875 | 0.4349 | 0.4816 | 0.4306 | 0.7222 |
| 1.4338 | 16.0 | 880 | 1.4030 | 0.4219 | 0.4630 | 0.4722 | 0.75 |
| 1.4338 | 17.0 | 935 | 1.3910 | 0.5924 | 0.5903 | 0.4722 | 0.8194 |
| 1.4338 | 18.0 | 990 | 1.3350 | 0.5842 | 0.5779 | 0.4861 | 0.8333 |
| 0.7075 | 19.0 | 1045 | 1.4847 | 0.5519 | 0.5815 | 0.375 | 0.7639 |
| 0.7075 | 20.0 | 1100 | 1.3782 | 0.5736 | 0.5788 | 0.4167 | 0.8056 |
Framework versions
- Transformers 4.26.0
- Pytorch 1.13.1
- Datasets 2.10.1
- Tokenizers 0.12.1