MiniLM_uncased_classification_tools_qlora_fr
This model is a fine-tuned version of microsoft/MiniLM-L12-H384-uncased on the None dataset. It achieves the following results on the evaluation set:
- Loss: 1.7495
- Accuracy: 0.5
- Learning Rate: 0.0
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: 0.0001
- train_batch_size: 24
- eval_batch_size: 192
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 60
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy | Rate |
|---|---|---|---|---|---|
| No log | 1.0 | 7 | 2.0845 | 0.075 | 0.0001 |
| No log | 2.0 | 14 | 2.0852 | 0.075 | 0.0001 |
| No log | 3.0 | 21 | 2.0851 | 0.075 | 0.0001 |
| No log | 4.0 | 28 | 2.0855 | 0.075 | 0.0001 |
| No log | 5.0 | 35 | 2.0858 | 0.075 | 0.0001 |
| No log | 6.0 | 42 | 2.0861 | 0.075 | 9e-05 |
| No log | 7.0 | 49 | 2.0863 | 0.075 | 0.0001 |
| No log | 8.0 | 56 | 2.0859 | 0.075 | 0.0001 |
| No log | 9.0 | 63 | 2.0856 | 0.075 | 0.0001 |
| No log | 10.0 | 70 | 2.0855 | 0.075 | 0.0001 |
| No log | 11.0 | 77 | 2.0848 | 0.075 | 0.0001 |
| No log | 12.0 | 84 | 2.0844 | 0.075 | 8e-05 |
| No log | 13.0 | 91 | 2.0829 | 0.075 | 0.0001 |
| No log | 14.0 | 98 | 2.0825 | 0.075 | 0.0001 |
| No log | 15.0 | 105 | 2.0811 | 0.1 | 0.0001 |
| No log | 16.0 | 112 | 2.0793 | 0.175 | 0.0001 |
| No log | 17.0 | 119 | 2.0771 | 0.125 | 0.0001 |
| No log | 18.0 | 126 | 2.0750 | 0.175 | 7e-05 |
| No log | 19.0 | 133 | 2.0723 | 0.175 | 0.0001 |
| No log | 20.0 | 140 | 2.0683 | 0.175 | 0.0001 |
| No log | 21.0 | 147 | 2.0637 | 0.175 | 0.0001 |
| No log | 22.0 | 154 | 2.0574 | 0.175 | 0.0001 |
| No log | 23.0 | 161 | 2.0507 | 0.2 | 0.0001 |
| No log | 24.0 | 168 | 2.0419 | 0.325 | 6e-05 |
| No log | 25.0 | 175 | 2.0318 | 0.35 | 0.0001 |
| No log | 26.0 | 182 | 2.0214 | 0.4 | 0.0001 |
| No log | 27.0 | 189 | 2.0083 | 0.4 | 0.0001 |
| No log | 28.0 | 196 | 1.9949 | 0.4 | 0.0001 |
| No log | 29.0 | 203 | 1.9781 | 0.4 | 0.0001 |
| No log | 30.0 | 210 | 1.9609 | 0.4 | 5e-05 |
| No log | 31.0 | 217 | 1.9475 | 0.425 | 0.0000 |
| No log | 32.0 | 224 | 1.9317 | 0.45 | 0.0000 |
| No log | 33.0 | 231 | 1.9131 | 0.45 | 0.0000 |
| No log | 34.0 | 238 | 1.9015 | 0.475 | 0.0000 |
| No log | 35.0 | 245 | 1.8906 | 0.5 | 0.0000 |
| No log | 36.0 | 252 | 1.8740 | 0.475 | 4e-05 |
| No log | 37.0 | 259 | 1.8613 | 0.5 | 0.0000 |
| No log | 38.0 | 266 | 1.8552 | 0.525 | 0.0000 |
| No log | 39.0 | 273 | 1.8389 | 0.5 | 0.0000 |
| No log | 40.0 | 280 | 1.8302 | 0.5 | 0.0000 |
| No log | 41.0 | 287 | 1.8228 | 0.5 | 0.0000 |
| No log | 42.0 | 294 | 1.8244 | 0.525 | 3e-05 |
| No log | 43.0 | 301 | 1.8048 | 0.5 | 0.0000 |
| No log | 44.0 | 308 | 1.7944 | 0.525 | 0.0000 |
| No log | 45.0 | 315 | 1.7929 | 0.5 | 0.0000 |
| No log | 46.0 | 322 | 1.7904 | 0.5 | 0.0000 |
| No log | 47.0 | 329 | 1.7810 | 0.5 | 0.0000 |
| No log | 48.0 | 336 | 1.7790 | 0.5 | 2e-05 |
| No log | 49.0 | 343 | 1.7758 | 0.5 | 0.0000 |
| No log | 50.0 | 350 | 1.7677 | 0.525 | 0.0000 |
| No log | 51.0 | 357 | 1.7626 | 0.525 | 0.0000 |
| No log | 52.0 | 364 | 1.7579 | 0.525 | 0.0000 |
| No log | 53.0 | 371 | 1.7552 | 0.525 | 0.0000 |
| No log | 54.0 | 378 | 1.7544 | 0.525 | 1e-05 |
| No log | 55.0 | 385 | 1.7523 | 0.525 | 0.0000 |
| No log | 56.0 | 392 | 1.7510 | 0.525 | 0.0000 |
| No log | 57.0 | 399 | 1.7501 | 0.525 | 5e-06 |
| No log | 58.0 | 406 | 1.7498 | 0.525 | 0.0000 |
| No log | 59.0 | 413 | 1.7496 | 0.525 | 0.0000 |
| No log | 60.0 | 420 | 1.7495 | 0.5 | 0.0 |
Framework versions
- Transformers 4.34.0
- Pytorch 2.0.1+cu117
- Datasets 2.14.5
- Tokenizers 0.14.1