metadata
tags:
- generated_from_trainer
model-index:
- name: GQA_RoBERTa_legal_SQuAD_complete_augmented_100
results: []
GQA_RoBERTa_legal_SQuAD_complete_augmented_100
This model was trained from scratch on the None dataset. It achieves the following results on the evaluation set:
- Loss: 3.7135
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: 128
- eval_batch_size: 32
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 100
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| No log | 1.0 | 4 | 5.0223 |
| No log | 2.0 | 8 | 2.8588 |
| No log | 3.0 | 12 | 3.7945 |
| No log | 4.0 | 16 | 2.6351 |
| No log | 5.0 | 20 | 3.0764 |
| No log | 6.0 | 24 | 2.4511 |
| No log | 7.0 | 28 | 2.5440 |
| No log | 8.0 | 32 | 2.8751 |
| No log | 9.0 | 36 | 2.6175 |
| No log | 10.0 | 40 | 2.9249 |
| No log | 11.0 | 44 | 2.3475 |
| No log | 12.0 | 48 | 2.8935 |
| No log | 13.0 | 52 | 2.7712 |
| No log | 14.0 | 56 | 2.5911 |
| No log | 15.0 | 60 | 3.1567 |
| No log | 16.0 | 64 | 2.7002 |
| No log | 17.0 | 68 | 3.0797 |
| No log | 18.0 | 72 | 2.9672 |
| No log | 19.0 | 76 | 3.2383 |
| No log | 20.0 | 80 | 3.1447 |
| No log | 21.0 | 84 | 3.1826 |
| No log | 22.0 | 88 | 3.4054 |
| No log | 23.0 | 92 | 3.1397 |
| No log | 24.0 | 96 | 3.3258 |
| No log | 25.0 | 100 | 3.2254 |
| No log | 26.0 | 104 | 3.1559 |
| No log | 27.0 | 108 | 3.5900 |
| No log | 28.0 | 112 | 3.3985 |
| No log | 29.0 | 116 | 3.7034 |
| No log | 30.0 | 120 | 3.6305 |
| No log | 31.0 | 124 | 3.4161 |
| No log | 32.0 | 128 | 3.6157 |
| No log | 33.0 | 132 | 3.3394 |
| No log | 34.0 | 136 | 3.5239 |
| No log | 35.0 | 140 | 3.4075 |
| No log | 36.0 | 144 | 3.3560 |
| No log | 37.0 | 148 | 3.6675 |
| No log | 38.0 | 152 | 3.3822 |
| No log | 39.0 | 156 | 3.5903 |
| No log | 40.0 | 160 | 3.7123 |
| No log | 41.0 | 164 | 3.5195 |
| No log | 42.0 | 168 | 3.6451 |
| No log | 43.0 | 172 | 3.4411 |
| No log | 44.0 | 176 | 3.6223 |
| No log | 45.0 | 180 | 3.4833 |
| No log | 46.0 | 184 | 3.7552 |
| No log | 47.0 | 188 | 3.5525 |
| No log | 48.0 | 192 | 3.7382 |
| No log | 49.0 | 196 | 3.6046 |
| No log | 50.0 | 200 | 3.6106 |
| No log | 51.0 | 204 | 3.6298 |
| No log | 52.0 | 208 | 3.7597 |
| No log | 53.0 | 212 | 3.5995 |
| No log | 54.0 | 216 | 3.6429 |
| No log | 55.0 | 220 | 3.6862 |
| No log | 56.0 | 224 | 3.6334 |
| No log | 57.0 | 228 | 4.0924 |
| No log | 58.0 | 232 | 3.6489 |
| No log | 59.0 | 236 | 3.6355 |
| No log | 60.0 | 240 | 3.8356 |
| No log | 61.0 | 244 | 3.5758 |
| No log | 62.0 | 248 | 3.5889 |
| No log | 63.0 | 252 | 3.7572 |
| No log | 64.0 | 256 | 3.7237 |
| No log | 65.0 | 260 | 3.6545 |
| No log | 66.0 | 264 | 3.7671 |
| No log | 67.0 | 268 | 3.6976 |
| No log | 68.0 | 272 | 3.6523 |
| No log | 69.0 | 276 | 3.7270 |
| No log | 70.0 | 280 | 3.7120 |
| No log | 71.0 | 284 | 3.6896 |
| No log | 72.0 | 288 | 3.6892 |
| No log | 73.0 | 292 | 3.6871 |
| No log | 74.0 | 296 | 3.7154 |
| No log | 75.0 | 300 | 3.6874 |
| No log | 76.0 | 304 | 3.7207 |
| No log | 77.0 | 308 | 3.7025 |
| No log | 78.0 | 312 | 3.6979 |
| No log | 79.0 | 316 | 3.7120 |
| No log | 80.0 | 320 | 3.7363 |
| No log | 81.0 | 324 | 3.7563 |
| No log | 82.0 | 328 | 3.7290 |
| No log | 83.0 | 332 | 3.6939 |
| No log | 84.0 | 336 | 3.6881 |
| No log | 85.0 | 340 | 3.6798 |
| No log | 86.0 | 344 | 3.6658 |
| No log | 87.0 | 348 | 3.6642 |
| No log | 88.0 | 352 | 3.6678 |
| No log | 89.0 | 356 | 3.6846 |
| No log | 90.0 | 360 | 3.6806 |
| No log | 91.0 | 364 | 3.6778 |
| No log | 92.0 | 368 | 3.6777 |
| No log | 93.0 | 372 | 3.6907 |
| No log | 94.0 | 376 | 3.7072 |
| No log | 95.0 | 380 | 3.7095 |
| No log | 96.0 | 384 | 3.7103 |
| No log | 97.0 | 388 | 3.7127 |
| No log | 98.0 | 392 | 3.7138 |
| No log | 99.0 | 396 | 3.7140 |
| No log | 100.0 | 400 | 3.7135 |
Framework versions
- Transformers 4.36.2
- Pytorch 2.1.2+cu121
- Datasets 2.14.7
- Tokenizers 0.15.0