whisper-small-sw-ndizi-158_2
This model is a fine-tuned version of pplantinga/whisper-small-sw on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 1.7989
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: 8
- 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 | 18 | 2.8008 |
| 2.5386 | 2.0 | 36 | 2.7557 |
| 2.4516 | 3.0 | 54 | 2.6801 |
| 2.4516 | 4.0 | 72 | 2.5842 |
| 2.363 | 5.0 | 90 | 2.5062 |
| 2.2145 | 6.0 | 108 | 2.4334 |
| 2.1358 | 7.0 | 126 | 2.3611 |
| 2.1358 | 8.0 | 144 | 2.2867 |
| 2.059 | 9.0 | 162 | 2.2158 |
| 1.8708 | 10.0 | 180 | 2.1452 |
| 1.8708 | 11.0 | 198 | 2.0739 |
| 1.8568 | 12.0 | 216 | 2.0307 |
| 1.8349 | 13.0 | 234 | 2.0092 |
| 1.7117 | 14.0 | 252 | 1.9914 |
| 1.7117 | 15.0 | 270 | 1.9788 |
| 1.6683 | 16.0 | 288 | 1.9669 |
| 1.6639 | 17.0 | 306 | 1.9548 |
| 1.6639 | 18.0 | 324 | 1.9432 |
| 1.6765 | 19.0 | 342 | 1.9338 |
| 1.6148 | 20.0 | 360 | 1.9265 |
| 1.5875 | 21.0 | 378 | 1.9149 |
| 1.5875 | 22.0 | 396 | 1.9058 |
| 1.5664 | 23.0 | 414 | 1.9013 |
| 1.6485 | 24.0 | 432 | 1.8953 |
| 1.535 | 25.0 | 450 | 1.8917 |
| 1.535 | 26.0 | 468 | 1.8849 |
| 1.5579 | 27.0 | 486 | 1.8783 |
| 1.5141 | 28.0 | 504 | 1.8746 |
| 1.5141 | 29.0 | 522 | 1.8707 |
| 1.5943 | 30.0 | 540 | 1.8663 |
| 1.4296 | 31.0 | 558 | 1.8630 |
| 1.4895 | 32.0 | 576 | 1.8595 |
| 1.4895 | 33.0 | 594 | 1.8536 |
| 1.5366 | 34.0 | 612 | 1.8525 |
| 1.4573 | 35.0 | 630 | 1.8488 |
| 1.4573 | 36.0 | 648 | 1.8489 |
| 1.4729 | 37.0 | 666 | 1.8441 |
| 1.4758 | 38.0 | 684 | 1.8384 |
| 1.4386 | 39.0 | 702 | 1.8391 |
| 1.4386 | 40.0 | 720 | 1.8356 |
| 1.3773 | 41.0 | 738 | 1.8343 |
| 1.4994 | 42.0 | 756 | 1.8328 |
| 1.4994 | 43.0 | 774 | 1.8331 |
| 1.4342 | 44.0 | 792 | 1.8287 |
| 1.4047 | 45.0 | 810 | 1.8283 |
| 1.3758 | 46.0 | 828 | 1.8261 |
| 1.3758 | 47.0 | 846 | 1.8234 |
| 1.3856 | 48.0 | 864 | 1.8201 |
| 1.3815 | 49.0 | 882 | 1.8210 |
| 1.4364 | 50.0 | 900 | 1.8197 |
| 1.4364 | 51.0 | 918 | 1.8188 |
| 1.4035 | 52.0 | 936 | 1.8183 |
| 1.3368 | 53.0 | 954 | 1.8176 |
| 1.3368 | 54.0 | 972 | 1.8155 |
| 1.424 | 55.0 | 990 | 1.8158 |
| 1.3782 | 56.0 | 1008 | 1.8159 |
| 1.3057 | 57.0 | 1026 | 1.8122 |
| 1.3057 | 58.0 | 1044 | 1.8136 |
| 1.3615 | 59.0 | 1062 | 1.8142 |
| 1.4013 | 60.0 | 1080 | 1.8091 |
| 1.4013 | 61.0 | 1098 | 1.8099 |
| 1.2894 | 62.0 | 1116 | 1.8102 |
| 1.3972 | 63.0 | 1134 | 1.8089 |
| 1.3564 | 64.0 | 1152 | 1.8096 |
| 1.3564 | 65.0 | 1170 | 1.8075 |
| 1.2808 | 66.0 | 1188 | 1.8078 |
| 1.3871 | 67.0 | 1206 | 1.8061 |
| 1.3871 | 68.0 | 1224 | 1.8060 |
| 1.267 | 69.0 | 1242 | 1.8070 |
| 1.2978 | 70.0 | 1260 | 1.8046 |
| 1.3657 | 71.0 | 1278 | 1.8062 |
| 1.3657 | 72.0 | 1296 | 1.8061 |
| 1.342 | 73.0 | 1314 | 1.8066 |
| 1.2504 | 74.0 | 1332 | 1.8063 |
| 1.3003 | 75.0 | 1350 | 1.8031 |
| 1.3003 | 76.0 | 1368 | 1.8053 |
| 1.2927 | 77.0 | 1386 | 1.8057 |
| 1.2653 | 78.0 | 1404 | 1.8032 |
| 1.2653 | 79.0 | 1422 | 1.8031 |
| 1.3574 | 80.0 | 1440 | 1.8036 |
| 1.2253 | 81.0 | 1458 | 1.8061 |
| 1.3348 | 82.0 | 1476 | 1.8036 |
| 1.3348 | 83.0 | 1494 | 1.8034 |
| 1.2846 | 84.0 | 1512 | 1.8033 |
| 1.2671 | 85.0 | 1530 | 1.8032 |
| 1.2671 | 86.0 | 1548 | 1.8038 |
| 1.3102 | 87.0 | 1566 | 1.8031 |
| 1.2603 | 88.0 | 1584 | 1.8011 |
| 1.286 | 89.0 | 1602 | 1.8029 |
| 1.286 | 90.0 | 1620 | 1.8026 |
| 1.2761 | 91.0 | 1638 | 1.8029 |
| 1.2416 | 92.0 | 1656 | 1.8014 |
| 1.2416 | 93.0 | 1674 | 1.8035 |
| 1.2798 | 94.0 | 1692 | 1.8008 |
| 1.3043 | 95.0 | 1710 | 1.8009 |
| 1.2969 | 96.0 | 1728 | 1.8004 |
| 1.2969 | 97.0 | 1746 | 1.8014 |
| 1.3087 | 98.0 | 1764 | 1.7992 |
| 1.2364 | 99.0 | 1782 | 1.7997 |
| 1.2748 | 100.0 | 1800 | 1.7989 |
Framework versions
- PEFT 0.7.2.dev0
- Transformers 4.37.0.dev0
- Pytorch 2.0.0
- Datasets 2.16.1
- Tokenizers 0.15.0