rule_learning_margin_1mm_many_negatives_spanpred_attention
This model is a fine-tuned version of enoriega/rule_softmatching on the enoriega/odinsynth_dataset dataset. It achieves the following results on the evaluation set:
- Loss: 0.2369
- Margin Accuracy: 0.8923
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-05
- train_batch_size: 4
- eval_batch_size: 4
- seed: 42
- gradient_accumulation_steps: 2000
- total_train_batch_size: 8000
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 3.0
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Margin Accuracy |
|---|---|---|---|---|
| 0.3814 | 0.16 | 20 | 0.3909 | 0.8317 |
| 0.349 | 0.32 | 40 | 0.3335 | 0.8463 |
| 0.3196 | 0.48 | 60 | 0.3101 | 0.8587 |
| 0.3083 | 0.64 | 80 | 0.3010 | 0.8645 |
| 0.2828 | 0.8 | 100 | 0.2871 | 0.8686 |
| 0.294 | 0.96 | 120 | 0.2800 | 0.8715 |
| 0.2711 | 1.12 | 140 | 0.2708 | 0.8741 |
| 0.2663 | 1.28 | 160 | 0.2671 | 0.8767 |
| 0.2656 | 1.44 | 180 | 0.2612 | 0.8822 |
| 0.2645 | 1.6 | 200 | 0.2537 | 0.8851 |
| 0.2625 | 1.76 | 220 | 0.2483 | 0.8878 |
| 0.2651 | 1.92 | 240 | 0.2471 | 0.8898 |
| 0.2407 | 2.08 | 260 | 0.2438 | 0.8905 |
| 0.2315 | 2.24 | 280 | 0.2408 | 0.8909 |
| 0.2461 | 2.4 | 300 | 0.2390 | 0.8918 |
| 0.2491 | 2.56 | 320 | 0.2390 | 0.8921 |
| 0.2511 | 2.72 | 340 | 0.2369 | 0.8918 |
| 0.2341 | 2.88 | 360 | 0.2363 | 0.8921 |
Framework versions
- Transformers 4.19.2
- Pytorch 1.11.0
- Datasets 2.2.1
- Tokenizers 0.12.1
- Downloads last month
- 3