whisper-small-sw-common-voice-ndizi-158-NF4
This model is a fine-tuned version of smutuvi/whisper-small-sw-common-voice on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 1.6482
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: 1e-05
- train_batch_size: 16
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 50
- num_epochs: 100
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| No log | 1.0 | 9 | 1.9783 |
| No log | 2.0 | 18 | 1.9757 |
| 1.8194 | 3.0 | 27 | 1.9710 |
| 1.8194 | 4.0 | 36 | 1.9631 |
| 1.8194 | 5.0 | 45 | 1.9543 |
| 1.7712 | 6.0 | 54 | 1.9437 |
| 1.7712 | 7.0 | 63 | 1.9337 |
| 1.7712 | 8.0 | 72 | 1.9244 |
| 1.7377 | 9.0 | 81 | 1.9141 |
| 1.7377 | 10.0 | 90 | 1.9043 |
| 1.7377 | 11.0 | 99 | 1.8952 |
| 1.7377 | 12.0 | 108 | 1.8863 |
| 1.7377 | 13.0 | 117 | 1.8762 |
| 1.6796 | 14.0 | 126 | 1.8684 |
| 1.6796 | 15.0 | 135 | 1.8588 |
| 1.6796 | 16.0 | 144 | 1.8501 |
| 1.6217 | 17.0 | 153 | 1.8414 |
| 1.6217 | 18.0 | 162 | 1.8325 |
| 1.6217 | 19.0 | 171 | 1.8253 |
| 1.6036 | 20.0 | 180 | 1.8169 |
| 1.6036 | 21.0 | 189 | 1.8097 |
| 1.6036 | 22.0 | 198 | 1.8017 |
| 1.6129 | 23.0 | 207 | 1.7949 |
| 1.6129 | 24.0 | 216 | 1.7879 |
| 1.5607 | 25.0 | 225 | 1.7817 |
| 1.5607 | 26.0 | 234 | 1.7760 |
| 1.5607 | 27.0 | 243 | 1.7699 |
| 1.5305 | 28.0 | 252 | 1.7637 |
| 1.5305 | 29.0 | 261 | 1.7588 |
| 1.5305 | 30.0 | 270 | 1.7537 |
| 1.5261 | 31.0 | 279 | 1.7485 |
| 1.5261 | 32.0 | 288 | 1.7444 |
| 1.5261 | 33.0 | 297 | 1.7400 |
| 1.5331 | 34.0 | 306 | 1.7360 |
| 1.5331 | 35.0 | 315 | 1.7322 |
| 1.5331 | 36.0 | 324 | 1.7287 |
| 1.5131 | 37.0 | 333 | 1.7250 |
| 1.5131 | 38.0 | 342 | 1.7216 |
| 1.4883 | 39.0 | 351 | 1.7183 |
| 1.4883 | 40.0 | 360 | 1.7151 |
| 1.4883 | 41.0 | 369 | 1.7119 |
| 1.4551 | 42.0 | 378 | 1.7092 |
| 1.4551 | 43.0 | 387 | 1.7063 |
| 1.4551 | 44.0 | 396 | 1.7042 |
| 1.47 | 45.0 | 405 | 1.7018 |
| 1.47 | 46.0 | 414 | 1.6988 |
| 1.47 | 47.0 | 423 | 1.6967 |
| 1.4422 | 48.0 | 432 | 1.6942 |
| 1.4422 | 49.0 | 441 | 1.6922 |
| 1.4432 | 50.0 | 450 | 1.6904 |
| 1.4432 | 51.0 | 459 | 1.6886 |
| 1.4432 | 52.0 | 468 | 1.6869 |
| 1.4108 | 53.0 | 477 | 1.6849 |
| 1.4108 | 54.0 | 486 | 1.6828 |
| 1.4108 | 55.0 | 495 | 1.6808 |
| 1.4377 | 56.0 | 504 | 1.6797 |
| 1.4377 | 57.0 | 513 | 1.6777 |
| 1.4377 | 58.0 | 522 | 1.6770 |
| 1.4281 | 59.0 | 531 | 1.6748 |
| 1.4281 | 60.0 | 540 | 1.6732 |
| 1.4281 | 61.0 | 549 | 1.6718 |
| 1.374 | 62.0 | 558 | 1.6704 |
| 1.374 | 63.0 | 567 | 1.6695 |
| 1.3947 | 64.0 | 576 | 1.6678 |
| 1.3947 | 65.0 | 585 | 1.6667 |
| 1.3947 | 66.0 | 594 | 1.6656 |
| 1.3843 | 67.0 | 603 | 1.6650 |
| 1.3843 | 68.0 | 612 | 1.6638 |
| 1.3843 | 69.0 | 621 | 1.6620 |
| 1.385 | 70.0 | 630 | 1.6614 |
| 1.385 | 71.0 | 639 | 1.6608 |
| 1.385 | 72.0 | 648 | 1.6592 |
| 1.368 | 73.0 | 657 | 1.6586 |
| 1.368 | 74.0 | 666 | 1.6576 |
| 1.3777 | 75.0 | 675 | 1.6574 |
| 1.3777 | 76.0 | 684 | 1.6565 |
| 1.3777 | 77.0 | 693 | 1.6556 |
| 1.3837 | 78.0 | 702 | 1.6553 |
| 1.3837 | 79.0 | 711 | 1.6547 |
| 1.3837 | 80.0 | 720 | 1.6544 |
| 1.3577 | 81.0 | 729 | 1.6529 |
| 1.3577 | 82.0 | 738 | 1.6523 |
| 1.3577 | 83.0 | 747 | 1.6515 |
| 1.3961 | 84.0 | 756 | 1.6513 |
| 1.3961 | 85.0 | 765 | 1.6512 |
| 1.3961 | 86.0 | 774 | 1.6505 |
| 1.3292 | 87.0 | 783 | 1.6497 |
| 1.3292 | 88.0 | 792 | 1.6495 |
| 1.3578 | 89.0 | 801 | 1.6496 |
| 1.3578 | 90.0 | 810 | 1.6495 |
| 1.3578 | 91.0 | 819 | 1.6488 |
| 1.3352 | 92.0 | 828 | 1.6486 |
| 1.3352 | 93.0 | 837 | 1.6492 |
| 1.3352 | 94.0 | 846 | 1.6487 |
| 1.3873 | 95.0 | 855 | 1.6485 |
| 1.3873 | 96.0 | 864 | 1.6483 |
| 1.3873 | 97.0 | 873 | 1.6481 |
| 1.3391 | 98.0 | 882 | 1.6483 |
| 1.3391 | 99.0 | 891 | 1.6482 |
| 1.3569 | 100.0 | 900 | 1.6482 |
Framework versions
- PEFT 0.7.2.dev0
- Transformers 4.37.0.dev0
- Pytorch 2.1.1+cu121
- Datasets 2.16.1
- Tokenizers 0.15.0