deberta-v3-large-survey-related_passage_consistency-rater-half
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: 0.7100
- Krippendorff: 0.5386
- Spearman: 0.5376
- Absolute Agreement: 0.7416
- Agreement Within One: 0.9062
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 | 52 | 2.0068 | -0.8382 | 0.2787 | 0.0 | 1.0 |
| No log | 2.0 | 104 | 1.9940 | -0.8214 | 0.2201 | 0.0 | 1.0 |
| No log | 3.0 | 156 | 1.9560 | -0.0763 | 0.2550 | 0.1667 | 0.9722 |
| No log | 4.0 | 208 | 2.0356 | -0.3464 | nan | 0.25 | 0.8889 |
| No log | 5.0 | 260 | 2.2692 | -0.3464 | nan | 0.25 | 0.8889 |
| No log | 6.0 | 312 | 2.0392 | -0.3464 | nan | 0.25 | 0.8889 |
| No log | 7.0 | 364 | 2.2373 | -0.3464 | nan | 0.25 | 0.8889 |
| No log | 8.0 | 416 | 2.1509 | -0.3464 | nan | 0.25 | 0.8889 |
| No log | 9.0 | 468 | 2.1995 | -0.3464 | nan | 0.25 | 0.8889 |
| 1.3238 | 10.0 | 520 | 2.2470 | -0.2831 | -0.2582 | 0.2222 | 0.8889 |
| 1.3238 | 11.0 | 572 | 2.6856 | -0.0211 | -0.0119 | 0.1944 | 0.9167 |
| 1.3238 | 12.0 | 624 | 2.8278 | -0.1707 | -0.2258 | 0.1111 | 0.9167 |
| 1.3238 | 13.0 | 676 | 2.8911 | -0.0941 | -0.2270 | 0.1667 | 0.9167 |
| 1.3238 | 14.0 | 728 | 2.5322 | -0.0560 | -0.1752 | 0.1667 | 0.9167 |
| 1.3238 | 15.0 | 780 | 2.6974 | -0.1705 | -0.1417 | 0.1389 | 0.9444 |
| 1.3238 | 16.0 | 832 | 3.1395 | -0.0871 | -0.1487 | 0.25 | 0.9167 |
| 1.3238 | 17.0 | 884 | 2.8212 | -0.1420 | -0.2250 | 0.2222 | 0.9444 |
| 1.3238 | 18.0 | 936 | 3.4453 | -0.1105 | -0.2704 | 0.2222 | 0.9167 |
| 1.3238 | 19.0 | 988 | 3.6874 | -0.1105 | -0.2704 | 0.2222 | 0.9167 |
| 0.5875 | 20.0 | 1040 | 4.2445 | -0.1094 | -0.2616 | 0.1944 | 0.9167 |
Framework versions
- Transformers 4.26.0
- Pytorch 1.13.1
- Datasets 2.10.1
- Tokenizers 0.12.1