Hugging Face's logo

golesheed
/
whisper-v2-Brabantic

Automatic Speech Recognition
Transformers
TensorBoard
Safetensors
Dutch
whisper
Generated from Trainer
Inference Endpoints
Model card Files Metrics Community

Whisper Large V2

This model is a fine-tuned version of openai/whisper-large-v2 on the None dataset. It achieves the following results on the evaluation set:

  • Loss: 0.3489
  • Wer: 17.3755

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: 3e-05
  • train_batch_size: 12
  • 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: 20
  • num_epochs: 5

Training results

Training Loss Epoch Step Validation Loss Wer
0.7517 0.0449 15 0.5024 40.5912
0.4299 0.0898 30 0.3868 34.1310
0.363 0.1347 45 0.3704 25.2001
0.3744 0.1796 60 0.3537 22.5218
0.3731 0.2246 75 0.3554 24.5736
0.3743 0.2695 90 0.3388 22.7566
0.3001 0.3144 105 0.3401 22.3597
0.3382 0.3593 120 0.3262 33.1556
0.3353 0.4042 135 0.3266 28.0469
0.325 0.4491 150 0.3247 26.7473
0.3303 0.4940 165 0.3147 22.8616
0.2925 0.5389 180 0.3147 21.2041
0.3109 0.5838 195 0.3108 23.1859
0.2989 0.6287 210 0.3084 24.3570
0.3111 0.6737 225 0.3018 18.4117
0.2918 0.7186 240 0.3033 17.6076
0.3099 0.7635 255 0.2971 21.7151
0.2997 0.8084 270 0.2987 21.5361
0.2898 0.8533 285 0.2923 21.5828
0.2848 0.8982 300 0.2914 17.6452
0.285 0.9431 315 0.2874 17.7425
0.2624 0.9880 330 0.2861 16.8489
0.169 1.0329 345 0.2948 18.5687
0.1515 1.0778 360 0.2927 26.6540
0.1504 1.1228 375 0.2918 18.9422
0.1484 1.1677 390 0.2916 18.3482
0.1358 1.2126 405 0.2904 17.2198
0.128 1.2575 420 0.2895 17.6764
0.1417 1.3024 435 0.2895 23.2572
0.1561 1.3473 450 0.2876 17.7775
0.1445 1.3922 465 0.2874 17.5415
0.1384 1.4371 480 0.2825 16.1420
0.1488 1.4820 495 0.2857 17.3832
0.1701 1.5269 510 0.2779 22.6826
0.1475 1.5719 525 0.2857 25.9860
0.144 1.6168 540 0.2790 16.3145
0.1402 1.6617 555 0.2874 21.3948
0.1575 1.7066 570 0.2756 15.9786
0.1409 1.7515 585 0.2815 17.0862
0.1388 1.7964 600 0.2792 18.9176
0.1273 1.8413 615 0.2803 23.6165
0.1537 1.8862 630 0.2758 17.5454
0.1537 1.9311 645 0.2764 15.8373
0.1474 1.9760 660 0.2708 16.4935
0.1111 2.0210 675 0.2805 19.4337
0.0745 2.0659 690 0.2924 18.5388
0.0639 2.1108 705 0.2917 15.8269
0.0673 2.1557 720 0.2945 16.9306
0.066 2.2006 735 0.2955 16.3677
0.0714 2.2455 750 0.2933 16.2289
0.0701 2.2904 765 0.2911 20.4558
0.0631 2.3353 780 0.2971 17.1316
0.064 2.3802 795 0.2916 15.3846
0.0659 2.4251 810 0.2971 15.1602
0.0615 2.4701 825 0.2878 20.4480
0.0723 2.5150 840 0.2935 14.7569
0.0695 2.5599 855 0.2846 15.6570
0.0704 2.6048 870 0.2919 19.4000
0.0642 2.6497 885 0.2849 17.7373
0.0684 2.6946 900 0.2888 15.9164
0.077 2.7395 915 0.2828 15.5052
0.0708 2.7844 930 0.2858 17.0538
0.065 2.8293 945 0.2829 20.8617
0.0788 2.8743 960 0.2854 19.5621
0.0677 2.9192 975 0.2825 16.6984
0.0642 2.9641 990 0.2887 16.1537
0.0627 3.0090 1005 0.2828 16.0331
0.0262 3.0539 1020 0.3084 15.0202
0.0266 3.0988 1035 0.3129 16.9708
0.024 3.1437 1050 0.3114 14.9722
0.0271 3.1886 1065 0.3152 14.5416
0.026 3.2335 1080 0.3135 16.4533
0.0281 3.2784 1095 0.3151 17.0123
0.0295 3.3234 1110 0.3160 15.4183
0.0259 3.3683 1125 0.3101 14.8269
0.0276 3.4132 1140 0.3194 14.1175
0.0271 3.4581 1155 0.3172 17.3314
0.0304 3.5030 1170 0.3111 18.0577
0.0268 3.5479 1185 0.3129 14.0928
0.0256 3.5928 1200 0.3083 14.7374
0.0281 3.6377 1215 0.3079 14.9125
0.0274 3.6826 1230 0.3180 14.4586
0.0282 3.7275 1245 0.3091 14.6622
0.0224 3.7725 1260 0.3139 14.4132
0.0254 3.8174 1275 0.3141 14.0747
0.0279 3.8623 1290 0.3110 18.3676
0.0245 3.9072 1305 0.3119 15.0565
0.0256 3.9521 1320 0.3149 16.3560
0.0273 3.9970 1335 0.3128 16.3405
0.0126 4.0419 1350 0.3265 14.9385
0.0087 4.0868 1365 0.3411 14.4547
0.009 4.1317 1380 0.3394 14.6298
0.0093 4.1766 1395 0.3424 14.4547
0.0082 4.2216 1410 0.3457 14.4780
0.0093 4.2665 1425 0.3472 13.8192
0.0072 4.3114 1440 0.3491 15.0189
0.0093 4.3563 1455 0.3490 16.3962
0.0098 4.4012 1470 0.3455 16.3755
0.0077 4.4461 1485 0.3429 16.9410
0.0089 4.4910 1500 0.3452 17.0966
0.0099 4.5359 1515 0.3469 18.3897
0.0066 4.5808 1530 0.3465 19.0083
0.0074 4.6257 1545 0.3455 19.6867
0.0069 4.6707 1560 0.3489 18.5440
0.008 4.7156 1575 0.3502 18.4078
0.0079 4.7605 1590 0.3503 18.1057
0.0077 4.8054 1605 0.3501 18.2574
0.0058 4.8503 1620 0.3492 18.1653
0.0076 4.8952 1635 0.3486 17.7905
0.0064 4.9401 1650 0.3487 17.3858
0.0057 4.9850 1665 0.3489 17.3755

Framework versions

  • Transformers 4.45.0.dev0
  • Pytorch 2.1.0+cu121
  • Datasets 2.20.0
  • Tokenizers 0.19.1
Downloads last month
10
Safetensors
Model size
1.54B params
Tensor type
F32
·
Inference Examples
Automatic Speech Recognition
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social visibility and check back later, or deploy to Inference Endpoints (dedicated) instead.

Model tree for golesheed/whisper-v2-Brabantic

Base model

openai/whisper-large-v2
Finetuned
(194)
this model

Evaluation results

Metadata error: specify a dataset to view leaderboard