whisper-small-sw-common-voice-ndizi-248-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: 1.2101
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 | 14 | 1.5885 |
| 1.7951 | 2.0 | 28 | 1.5840 |
| 1.7951 | 3.0 | 42 | 1.5757 |
| 1.7624 | 4.0 | 56 | 1.5638 |
| 1.7624 | 5.0 | 70 | 1.5518 |
| 1.7014 | 6.0 | 84 | 1.5391 |
| 1.7014 | 7.0 | 98 | 1.5258 |
| 1.7576 | 8.0 | 112 | 1.5133 |
| 1.6769 | 9.0 | 126 | 1.5003 |
| 1.6769 | 10.0 | 140 | 1.4872 |
| 1.6555 | 11.0 | 154 | 1.4748 |
| 1.6555 | 12.0 | 168 | 1.4620 |
| 1.6441 | 13.0 | 182 | 1.4507 |
| 1.6441 | 14.0 | 196 | 1.4393 |
| 1.5213 | 15.0 | 210 | 1.4289 |
| 1.5213 | 16.0 | 224 | 1.4184 |
| 1.5634 | 17.0 | 238 | 1.4092 |
| 1.585 | 18.0 | 252 | 1.4002 |
| 1.585 | 19.0 | 266 | 1.3918 |
| 1.4976 | 20.0 | 280 | 1.3842 |
| 1.4976 | 21.0 | 294 | 1.3772 |
| 1.5197 | 22.0 | 308 | 1.3701 |
| 1.5197 | 23.0 | 322 | 1.3634 |
| 1.501 | 24.0 | 336 | 1.3572 |
| 1.4607 | 25.0 | 350 | 1.3513 |
| 1.4607 | 26.0 | 364 | 1.3459 |
| 1.4886 | 27.0 | 378 | 1.3407 |
| 1.4886 | 28.0 | 392 | 1.3358 |
| 1.4133 | 29.0 | 406 | 1.3305 |
| 1.4133 | 30.0 | 420 | 1.3262 |
| 1.4263 | 31.0 | 434 | 1.3216 |
| 1.4263 | 32.0 | 448 | 1.3173 |
| 1.4365 | 33.0 | 462 | 1.3126 |
| 1.42 | 34.0 | 476 | 1.3090 |
| 1.42 | 35.0 | 490 | 1.3053 |
| 1.4301 | 36.0 | 504 | 1.3015 |
| 1.4301 | 37.0 | 518 | 1.2983 |
| 1.3565 | 38.0 | 532 | 1.2949 |
| 1.3565 | 39.0 | 546 | 1.2919 |
| 1.4627 | 40.0 | 560 | 1.2887 |
| 1.4627 | 41.0 | 574 | 1.2854 |
| 1.3976 | 42.0 | 588 | 1.2828 |
| 1.3613 | 43.0 | 602 | 1.2794 |
| 1.3613 | 44.0 | 616 | 1.2770 |
| 1.3281 | 45.0 | 630 | 1.2739 |
| 1.3281 | 46.0 | 644 | 1.2716 |
| 1.352 | 47.0 | 658 | 1.2689 |
| 1.352 | 48.0 | 672 | 1.2662 |
| 1.3798 | 49.0 | 686 | 1.2637 |
| 1.3523 | 50.0 | 700 | 1.2617 |
| 1.3523 | 51.0 | 714 | 1.2592 |
| 1.369 | 52.0 | 728 | 1.2573 |
| 1.369 | 53.0 | 742 | 1.2549 |
| 1.3467 | 54.0 | 756 | 1.2528 |
| 1.3467 | 55.0 | 770 | 1.2510 |
| 1.3398 | 56.0 | 784 | 1.2488 |
| 1.3398 | 57.0 | 798 | 1.2470 |
| 1.2938 | 58.0 | 812 | 1.2453 |
| 1.3292 | 59.0 | 826 | 1.2433 |
| 1.3292 | 60.0 | 840 | 1.2418 |
| 1.3721 | 61.0 | 854 | 1.2402 |
| 1.3721 | 62.0 | 868 | 1.2383 |
| 1.2928 | 63.0 | 882 | 1.2369 |
| 1.2928 | 64.0 | 896 | 1.2353 |
| 1.3214 | 65.0 | 910 | 1.2336 |
| 1.3214 | 66.0 | 924 | 1.2321 |
| 1.3324 | 67.0 | 938 | 1.2308 |
| 1.2997 | 68.0 | 952 | 1.2296 |
| 1.2997 | 69.0 | 966 | 1.2282 |
| 1.3038 | 70.0 | 980 | 1.2271 |
| 1.3038 | 71.0 | 994 | 1.2258 |
| 1.3161 | 72.0 | 1008 | 1.2248 |
| 1.3161 | 73.0 | 1022 | 1.2240 |
| 1.2948 | 74.0 | 1036 | 1.2228 |
| 1.2907 | 75.0 | 1050 | 1.2220 |
| 1.2907 | 76.0 | 1064 | 1.2211 |
| 1.3348 | 77.0 | 1078 | 1.2200 |
| 1.3348 | 78.0 | 1092 | 1.2190 |
| 1.3043 | 79.0 | 1106 | 1.2181 |
| 1.3043 | 80.0 | 1120 | 1.2177 |
| 1.2185 | 81.0 | 1134 | 1.2167 |
| 1.2185 | 82.0 | 1148 | 1.2160 |
| 1.3256 | 83.0 | 1162 | 1.2151 |
| 1.3329 | 84.0 | 1176 | 1.2146 |
| 1.3329 | 85.0 | 1190 | 1.2142 |
| 1.263 | 86.0 | 1204 | 1.2134 |
| 1.263 | 87.0 | 1218 | 1.2131 |
| 1.3019 | 88.0 | 1232 | 1.2126 |
| 1.3019 | 89.0 | 1246 | 1.2120 |
| 1.2682 | 90.0 | 1260 | 1.2117 |
| 1.2682 | 91.0 | 1274 | 1.2113 |
| 1.323 | 92.0 | 1288 | 1.2113 |
| 1.2818 | 93.0 | 1302 | 1.2109 |
| 1.2818 | 94.0 | 1316 | 1.2107 |
| 1.3163 | 95.0 | 1330 | 1.2106 |
| 1.3163 | 96.0 | 1344 | 1.2103 |
| 1.2391 | 97.0 | 1358 | 1.2101 |
| 1.2391 | 98.0 | 1372 | 1.2103 |
| 1.2595 | 99.0 | 1386 | 1.2101 |
| 1.283 | 100.0 | 1400 | 1.2101 |
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