jlvdoorn's picture
Update README.md
64ed143
|
raw
history blame
4.56 kB
metadata
license: apache-2.0
tags:
  - generated_from_trainer
metrics:
  - wer
model-index:
  - name: whisper-large-v2-atcosim
    results: []

whisper-large-v2-atcosim

This model is a fine-tuned version of openai/whisper-large-v2 on the ATCOSIM dataset. It achieves the following results on the evaluation set:

  • Loss: 0.0679
  • Wer: 6.2234

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
  • distributed_type: multi-GPU
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • lr_scheduler_warmup_steps: 1000
  • training_steps: 50000

Training results

Training Loss Epoch Step Validation Loss Wer
0.0147 2.09 1000 0.0373 4.7972
0.0058 4.18 2000 0.0379 4.5934
0.0012 6.28 3000 0.0388 4.5425
0.0067 8.37 4000 0.0382 4.0470
0.0026 10.46 5000 0.0382 2.9959
0.0016 12.55 6000 0.0457 2.9496
0.0027 14.64 7000 0.0473 4.6305
0.001 16.74 8000 0.0419 3.2969
0.0011 18.83 9000 0.0424 4.4592
0.0013 20.92 10000 0.0432 5.8807
0.0019 23.01 11000 0.0454 3.7646
0.0004 25.1 12000 0.0443 9.5110
0.004 27.2 13000 0.0442 2.8385
0.0018 29.29 14000 0.0444 2.5282
0.0011 31.38 15000 0.0467 4.0980
0.0002 33.47 16000 0.0469 3.9128
0.003 35.56 17000 0.0454 4.7462
0.0001 37.66 18000 0.0459 3.1950
0.0006 39.75 19000 0.0451 2.6579
0.0014 41.84 20000 0.0464 1.6855
0.0 43.93 21000 0.0487 2.3106
0.0005 46.03 22000 0.0535 7.3717
0.0001 48.12 23000 0.0482 6.9411
0.0002 50.21 24000 0.0484 13.0580
0.0001 52.3 25000 0.0481 18.0219
0.0 54.39 26000 0.0523 14.7342
0.0 56.49 27000 0.0552 11.1132
0.0004 58.58 28000 0.0521 2.5190
0.0001 60.67 29000 0.0490 4.4036
0.0 62.76 30000 0.0497 2.8246
0.0 64.85 31000 0.0513 2.8755
0.0 66.95 32000 0.0526 2.9172
0.0 69.04 33000 0.0539 3.0098
0.0 71.13 34000 0.0552 3.0144
0.0 73.22 35000 0.0566 3.1209
0.0 75.31 36000 0.0580 3.2321
0.0 77.41 37000 0.0594 3.4729
0.0 79.5 38000 0.0607 3.6164
0.0 81.59 39000 0.0620 3.9035
0.0 83.68 40000 0.0632 4.0656
0.0 85.77 41000 0.0642 4.3202
0.0 87.87 42000 0.0651 4.4453
0.0 89.96 43000 0.0659 4.9361
0.0 92.05 44000 0.0664 5.2186
0.0 94.14 45000 0.0670 5.6029
0.0 96.23 46000 0.0673 5.7835
0.0 98.33 47000 0.0676 6.0520
0.0 100.42 48000 0.0678 6.1122
0.0 102.51 49000 0.0679 6.2141
0.0 104.6 50000 0.0679 6.2234

Framework versions

  • Transformers 4.30.0.dev0
  • Pytorch 2.0.1+cu117
  • Datasets 2.12.0
  • Tokenizers 0.13.3