metadata
tags:
- generated_from_trainer
datasets:
- enoriega/odinsynth_dataset
model-index:
- name: rule_learning_margin_1mm_spanpred_attention
results: []
rule_learning_margin_1mm_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: 0.3237
- Margin Accuracy: 0.8518
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 |
|---|---|---|---|---|
| 0.5768 | 0.16 | 20 | 0.5693 | 0.7577 |
| 0.4593 | 0.32 | 40 | 0.4338 | 0.8105 |
| 0.4219 | 0.48 | 60 | 0.3958 | 0.8218 |
| 0.3953 | 0.64 | 80 | 0.3809 | 0.8308 |
| 0.383 | 0.8 | 100 | 0.3684 | 0.8355 |
| 0.3781 | 0.96 | 120 | 0.3591 | 0.8396 |
| 0.354 | 1.12 | 140 | 0.3535 | 0.8420 |
| 0.3521 | 1.28 | 160 | 0.3491 | 0.8430 |
| 0.3533 | 1.44 | 180 | 0.3423 | 0.8466 |
| 0.344 | 1.6 | 200 | 0.3372 | 0.8472 |
| 0.3352 | 1.76 | 220 | 0.3345 | 0.8478 |
| 0.3318 | 1.92 | 240 | 0.3320 | 0.8487 |
| 0.3478 | 2.08 | 260 | 0.3286 | 0.8494 |
| 0.3329 | 2.24 | 280 | 0.3286 | 0.8505 |
| 0.3424 | 2.4 | 300 | 0.3262 | 0.8506 |
| 0.3463 | 2.56 | 320 | 0.3264 | 0.8512 |
| 0.3416 | 2.72 | 340 | 0.3247 | 0.8518 |
| 0.329 | 2.88 | 360 | 0.3247 | 0.8516 |
Framework versions
- Transformers 4.19.2
- Pytorch 1.11.0
- Datasets 2.2.1
- Tokenizers 0.12.1