Hugging Face's logo

cantillation
/
Teamim-small_WeightDecay-0.05_Augmented_New-Data_date-19-07-2024_15-41

Automatic Speech Recognition
Transformers
TensorBoard
Safetensors
Hebrew
whisper
hf-asr-leaderboard
Generated from Trainer
Inference Endpoints
Model card Files Metrics Community
Teamim-small_WeightDecay-0.05_Augmented_New-Data_date-19-07-2024_15-41 / README.md
cantillation's picture
cantillation
End of training
5142dd8 verified
preview code
|
raw
history blame contribute delete
19.1 kB
metadata
language:
  - he
license: apache-2.0
base_model: openai/whisper-small
tags:
  - hf-asr-leaderboard
  - generated_from_trainer
metrics:
  - wer
model-index:
  - name: he-cantillation
    results: []

he-cantillation

This model is a fine-tuned version of openai/whisper-small on an unknown dataset. It achieves the following results on the evaluation set:

  • Loss: 0.2357
  • Wer: 12.4316
  • Avg Precision Exact: 0.8908
  • Avg Recall Exact: 0.8920
  • Avg F1 Exact: 0.8911
  • Avg Precision Letter Shift: 0.9098
  • Avg Recall Letter Shift: 0.9110
  • Avg F1 Letter Shift: 0.9100
  • Avg Precision Word Level: 0.9122
  • Avg Recall Word Level: 0.9134
  • Avg F1 Word Level: 0.9125
  • Avg Precision Word Shift: 0.9710
  • Avg Recall Word Shift: 0.9730
  • Avg F1 Word Shift: 0.9716
  • Precision Median Exact: 1.0
  • Recall Median Exact: 1.0
  • F1 Median Exact: 1.0
  • Precision Max Exact: 1.0
  • Recall Max Exact: 1.0
  • F1 Max Exact: 1.0
  • Precision Min Exact: 0.0
  • Recall Min Exact: 0.0
  • F1 Min Exact: 0.0
  • Precision Min Letter Shift: 0.0
  • Recall Min Letter Shift: 0.0
  • F1 Min Letter Shift: 0.0
  • Precision Min Word Level: 0.0
  • Recall Min Word Level: 0.0
  • F1 Min Word Level: 0.0
  • Precision Min Word Shift: 0.1429
  • Recall Min Word Shift: 0.1111
  • F1 Min Word Shift: 0.125

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: 32
  • seed: 42
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • lr_scheduler_warmup_steps: 500
  • training_steps: 200000
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Validation Loss Wer Avg Precision Exact Avg Recall Exact Avg F1 Exact Avg Precision Letter Shift Avg Recall Letter Shift Avg F1 Letter Shift Avg Precision Word Level Avg Recall Word Level Avg F1 Word Level Avg Precision Word Shift Avg Recall Word Shift Avg F1 Word Shift Precision Median Exact Recall Median Exact F1 Median Exact Precision Max Exact Recall Max Exact F1 Max Exact Precision Min Exact Recall Min Exact F1 Min Exact Precision Min Letter Shift Recall Min Letter Shift F1 Min Letter Shift Precision Min Word Level Recall Min Word Level F1 Min Word Level Precision Min Word Shift Recall Min Word Shift F1 Min Word Shift
No log 0.0001 1 8.6313 200.3067 0.0004 0.0014 0.0004 0.0046 0.0050 0.0041 0.0031 0.0092 0.0033 0.0326 0.0360 0.0303 0.0 0.0 0.0 0.1111 1.0 0.2000 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
0.0158 0.9410 10000 0.1501 17.4538 0.8417 0.8453 0.8428 0.8645 0.8682 0.8656 0.8682 0.8723 0.8695 0.9449 0.9497 0.9463 0.9231 0.9231 0.9231 1.0 1.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0909 0.0909 0.1111
0.0149 1.8820 20000 0.1635 15.7724 0.8627 0.8659 0.8639 0.8851 0.8885 0.8864 0.8881 0.8915 0.8894 0.9560 0.9594 0.9571 0.9286 0.9333 0.9412 1.0 1.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
0.0053 2.8230 30000 0.1755 14.9261 0.8767 0.8763 0.8761 0.8990 0.8986 0.8983 0.9016 0.9013 0.9010 0.9609 0.9621 0.9610 0.9333 0.9375 0.9474 1.0 1.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.1429 0.125 0.1333
0.0034 3.7640 40000 0.1817 15.5691 0.8584 0.8592 0.8584 0.8794 0.8804 0.8794 0.8823 0.8833 0.8824 0.9538 0.9566 0.9547 0.9333 0.9333 0.9444 1.0 1.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
0.0005 4.7050 50000 0.1857 14.4050 0.8793 0.8793 0.8789 0.8998 0.8999 0.8994 0.9024 0.9028 0.9022 0.9629 0.9637 0.9628 0.9474 1.0 0.9565 1.0 1.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.1429 0.1 0.1176
0.0021 5.6460 60000 0.1945 14.5196 0.8716 0.8762 0.8734 0.8930 0.8978 0.8949 0.8960 0.9006 0.8979 0.9607 0.9659 0.9628 0.9333 1.0 0.9565 1.0 1.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0769 0.1 0.0952
0.0016 6.5870 70000 0.1987 14.0872 0.8773 0.8781 0.8772 0.8985 0.8993 0.8984 0.9015 0.9025 0.9015 0.9641 0.9661 0.9646 0.9412 1.0 0.9600 1.0 1.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.1429 0.1111 0.125
0.0005 7.5280 80000 0.2020 13.7288 0.8801 0.8825 0.8809 0.9015 0.9041 0.9024 0.9044 0.9069 0.9052 0.9673 0.9709 0.9687 1.0 1.0 0.9630 1.0 1.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.1429 0.1 0.1176
0.0021 8.4690 90000 0.2088 14.3829 0.8767 0.8782 0.8770 0.8980 0.8997 0.8984 0.9009 0.9026 0.9013 0.9629 0.9664 0.9640 0.9412 1.0 0.9600 1.0 1.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
0.0013 9.4100 100000 0.2041 13.4146 0.8826 0.8837 0.8828 0.9025 0.9037 0.9027 0.9052 0.9063 0.9054 0.9681 0.9697 0.9684 1.0 1.0 0.9630 1.0 1.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.1429 0.1111 0.125
0.0002 10.3510 110000 0.2110 13.5329 0.8876 0.8891 0.8880 0.9086 0.9101 0.9089 0.9112 0.9127 0.9115 0.9681 0.9710 0.9690 1.0 1.0 0.9630 1.0 1.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.1429 0.1 0.1176
0.0003 11.2920 120000 0.2123 13.5033 0.8822 0.8828 0.8821 0.9023 0.9029 0.9022 0.9048 0.9056 0.9048 0.9679 0.9698 0.9684 1.0 1.0 0.9677 1.0 1.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.1429 0.1 0.1176
0.0011 12.2330 130000 0.2082 13.5070 0.8884 0.8881 0.8878 0.9085 0.9083 0.9080 0.9109 0.9107 0.9104 0.9677 0.9683 0.9675 1.0 1.0 0.9677 1.0 1.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0769 0.0769 0.0769
0.0 13.1740 140000 0.2181 13.4220 0.8835 0.8860 0.8844 0.9028 0.9054 0.9038 0.9056 0.9082 0.9065 0.9671 0.9706 0.9684 1.0 1.0 0.9677 1.0 1.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.1429 0.1111 0.125
0.0 14.1150 150000 0.2240 13.0266 0.8847 0.8860 0.8850 0.9043 0.9057 0.9046 0.9073 0.9086 0.9075 0.9692 0.9719 0.9701 1.0 1.0 0.9677 1.0 1.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.1429 0.1111 0.125
0.0002 15.0560 160000 0.2245 13.0155 0.8828 0.8831 0.8826 0.9019 0.9022 0.9017 0.9049 0.9053 0.9048 0.9701 0.9717 0.9704 1.0 1.0 0.9687 1.0 1.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.1429 0.1111 0.125
0.0001 15.9970 170000 0.2272 13.0155 0.8855 0.8861 0.8854 0.9043 0.9049 0.9042 0.9070 0.9078 0.9070 0.9693 0.9710 0.9697 1.0 1.0 0.9677 1.0 1.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.1429 0.1111 0.125
0.0 16.9380 180000 0.2312 12.6053 0.8875 0.8892 0.8880 0.9062 0.9079 0.9067 0.9089 0.9106 0.9093 0.9703 0.9728 0.9711 1.0 1.0 1.0 1.0 1.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.1429 0.1111 0.125
0.0 17.8790 190000 0.2324 12.5684 0.8898 0.8910 0.8900 0.9089 0.9102 0.9092 0.9113 0.9127 0.9116 0.9703 0.9726 0.9710 1.0 1.0 1.0 1.0 1.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.1429 0.1111 0.125
0.0 18.8200 200000 0.2357 12.4316 0.8908 0.8920 0.8911 0.9098 0.9110 0.9100 0.9122 0.9134 0.9125 0.9710 0.9730 0.9716 1.0 1.0 1.0 1.0 1.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.1429 0.1111 0.125

Framework versions

  • Transformers 4.41.2
  • Pytorch 2.2.1
  • Datasets 2.20.0
  • Tokenizers 0.19.1