Configuration Parsing
Warning:
In adapter_config.json: "peft.task_type" must be a string
Whisper Turbo ko
This model is a fine-tuned version of openai/whisper-large-v3-turbo on the custom dataset. It achieves the following results on the evaluation set:
- Loss: 0.0725
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: 0.001
- train_batch_size: 64
- eval_batch_size: 256
- seed: 42
- optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 200
- training_steps: 1000
- mixed_precision_training: Native AMP
Training results
Training Loss | Epoch | Step | Validation Loss |
---|---|---|---|
0.7644 | 0.0625 | 10 | 1.3650 |
0.6675 | 0.125 | 20 | 1.1146 |
0.3476 | 0.1875 | 30 | 0.6689 |
0.2057 | 0.25 | 40 | 0.5514 |
0.1482 | 0.3125 | 50 | 0.4671 |
0.1417 | 0.375 | 60 | 0.4373 |
0.1271 | 0.4375 | 70 | 0.3908 |
0.1216 | 0.5 | 80 | 0.3624 |
0.0983 | 0.5625 | 90 | 0.3465 |
0.0966 | 0.625 | 100 | 0.3262 |
0.0907 | 0.6875 | 110 | 0.3118 |
0.0909 | 0.75 | 120 | 0.2852 |
0.0903 | 0.8125 | 130 | 0.2766 |
0.0783 | 0.875 | 140 | 0.2729 |
0.0761 | 0.9375 | 150 | 0.2519 |
0.074 | 1.0 | 160 | 0.2351 |
0.0675 | 1.0625 | 170 | 0.2196 |
0.0784 | 1.125 | 180 | 0.1990 |
0.057 | 1.1875 | 190 | 0.1988 |
0.0577 | 1.25 | 200 | 0.2089 |
0.0575 | 1.3125 | 210 | 0.2039 |
0.0653 | 1.375 | 220 | 0.1941 |
0.0681 | 1.4375 | 230 | 0.1894 |
0.0923 | 1.5 | 240 | 0.1880 |
0.0678 | 1.5625 | 250 | 0.1892 |
0.0685 | 1.625 | 260 | 0.1796 |
0.0643 | 1.6875 | 270 | 0.1687 |
0.0654 | 1.75 | 280 | 0.1668 |
0.0672 | 1.8125 | 290 | 0.1721 |
0.0692 | 1.875 | 300 | 0.1631 |
0.0728 | 1.9375 | 310 | 0.1602 |
0.0808 | 2.0 | 320 | 0.1884 |
0.0583 | 2.0625 | 330 | 0.1891 |
0.0531 | 2.125 | 340 | 0.1755 |
0.0647 | 2.1875 | 350 | 0.1793 |
0.0525 | 2.25 | 360 | 0.1651 |
0.0591 | 2.3125 | 370 | 0.1585 |
0.0488 | 2.375 | 380 | 0.1495 |
0.042 | 2.4375 | 390 | 0.1461 |
0.0491 | 2.5 | 400 | 0.1323 |
0.0472 | 2.5625 | 410 | 0.1390 |
0.05 | 2.625 | 420 | 0.1395 |
0.0377 | 2.6875 | 430 | 0.1424 |
0.0743 | 2.75 | 440 | 0.1334 |
0.0501 | 2.8125 | 450 | 0.1534 |
0.052 | 2.875 | 460 | 0.1567 |
0.0474 | 2.9375 | 470 | 0.1433 |
0.0545 | 3.0 | 480 | 0.1356 |
0.0296 | 3.0625 | 490 | 0.1285 |
0.0255 | 3.125 | 500 | 0.1257 |
0.0361 | 3.1875 | 510 | 0.1216 |
0.0355 | 3.25 | 520 | 0.1194 |
0.0294 | 3.3125 | 530 | 0.1193 |
0.0251 | 3.375 | 540 | 0.1170 |
0.0331 | 3.4375 | 550 | 0.1151 |
0.0322 | 3.5 | 560 | 0.1110 |
0.0347 | 3.5625 | 570 | 0.1105 |
0.0671 | 3.625 | 580 | 0.1559 |
0.0323 | 3.6875 | 590 | 0.1433 |
0.0349 | 3.75 | 600 | 0.1444 |
0.0344 | 3.8125 | 610 | 0.1398 |
0.0454 | 3.875 | 620 | 0.1438 |
0.0282 | 3.9375 | 630 | 0.1427 |
0.0323 | 4.0 | 640 | 0.1403 |
0.0258 | 4.0625 | 650 | 0.1361 |
0.0208 | 4.125 | 660 | 0.1350 |
0.0218 | 4.1875 | 670 | 0.1325 |
0.0174 | 4.25 | 680 | 0.1394 |
0.0238 | 4.3125 | 690 | 0.1333 |
0.0186 | 4.375 | 700 | 0.1335 |
0.0226 | 4.4375 | 710 | 0.1337 |
0.0229 | 4.5 | 720 | 0.1314 |
0.0234 | 4.5625 | 730 | 0.1289 |
0.0185 | 4.625 | 740 | 0.1254 |
0.086 | 4.6875 | 750 | 0.0999 |
0.0195 | 4.75 | 760 | 0.1028 |
0.02 | 4.8125 | 770 | 0.1015 |
0.0199 | 4.875 | 780 | 0.1024 |
0.0258 | 4.9375 | 790 | 0.0969 |
0.0196 | 5.0 | 800 | 0.0956 |
0.0145 | 5.0625 | 810 | 0.0925 |
0.0132 | 5.125 | 820 | 0.0924 |
0.0146 | 5.1875 | 830 | 0.0909 |
0.0124 | 5.25 | 840 | 0.0902 |
0.0151 | 5.3125 | 850 | 0.0899 |
0.0142 | 5.375 | 860 | 0.0890 |
0.0175 | 5.4375 | 870 | 0.0888 |
0.0596 | 5.5 | 880 | 0.0778 |
0.0147 | 5.5625 | 890 | 0.0768 |
0.0178 | 5.625 | 900 | 0.0759 |
0.0144 | 5.6875 | 910 | 0.0759 |
0.0131 | 5.75 | 920 | 0.0754 |
0.0114 | 5.8125 | 930 | 0.0742 |
0.0145 | 5.875 | 940 | 0.0735 |
0.0202 | 5.9375 | 950 | 0.0734 |
0.0146 | 6.0 | 960 | 0.0734 |
0.0109 | 6.0625 | 970 | 0.0729 |
0.0102 | 6.125 | 980 | 0.0727 |
0.0125 | 6.1875 | 990 | 0.0726 |
0.0099 | 6.25 | 1000 | 0.0725 |
Framework versions
- PEFT 0.14.0
- Transformers 4.47.1
- Pytorch 2.5.1+cu124
- Datasets 3.2.0
- Tokenizers 0.21.0
- Downloads last month
- 0
Inference Providers
NEW
This model is not currently available via any of the supported Inference Providers.
The model cannot be deployed to the HF Inference API:
The model has no pipeline_tag.