Hugging Face's logo

cantillation
/
whisper-tiny-he-teamim-allNusah-14-03-24-warmup-100-RandomFalse

Automatic Speech Recognition
Transformers
TensorBoard
Safetensors
Hebrew
whisper
hf-asr-leaderboard
Generated from Trainer
Model card Files Metrics Community
whisper-tiny-he-teamim-allNusah-14-03-24-warmup-100-RandomFalse / README.md
cantillation's picture
cantillation
End of training
6933e48 verified
preview code
|
raw
history blame
14.6 kB
metadata
language:
  - he
license: apache-2.0
base_model: openai/whisper-medium
tags:
  - hf-asr-leaderboard
  - generated_from_trainer
metrics:
  - wer
model-index:
  - name: he
    results: []

he

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

  • Loss: 0.1692
  • Wer: 21.1641
  • Avg Precision Exact: 0.8062
  • Avg Recall Exact: 0.8080
  • Avg F1 Exact: 0.8065
  • Avg Precision Letter Shift: 0.8352
  • Avg Recall Letter Shift: 0.8370
  • Avg F1 Letter Shift: 0.8355
  • Avg Precision Word Level: 0.8409
  • Avg Recall Word Level: 0.8430
  • Avg F1 Word Level: 0.8413
  • Avg Precision Word Shift: 0.9407
  • Avg Recall Word Shift: 0.9446
  • Avg F1 Word Shift: 0.9419
  • Precision Median Exact: 0.9091
  • Recall Median Exact: 0.9091
  • F1 Median Exact: 0.9091
  • 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: 0.1
  • F1 Min Word Shift: 0.1176

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: 100
  • training_steps: 30000
  • 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
0.2936 0.16 2000 0.3109 44.8004 0.5576 0.5646 0.5602 0.6021 0.6096 0.6049 0.6142 0.6239 0.6181 0.8035 0.8191 0.8099 0.6154 0.625 0.6207 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
0.1696 0.32 4000 0.2363 35.3622 0.6570 0.6509 0.6531 0.6965 0.6898 0.6923 0.7059 0.7000 0.7021 0.8774 0.8761 0.8755 0.75 0.75 0.75 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
0.1147 0.48 6000 0.2059 30.2846 0.7069 0.7015 0.7035 0.7422 0.7364 0.7385 0.7512 0.7461 0.7478 0.8992 0.8989 0.8980 0.8182 0.8 0.8000 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
0.101 0.64 8000 0.1887 27.1175 0.7477 0.7487 0.7475 0.7812 0.7822 0.7810 0.7884 0.7898 0.7884 0.9171 0.9216 0.9183 0.8462 0.8571 0.8571 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
0.0625 0.8 10000 0.1856 25.9313 0.7457 0.7570 0.7506 0.7776 0.7894 0.7827 0.7838 0.7952 0.7888 0.9171 0.9287 0.9219 0.8571 0.875 0.8571 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
0.0744 0.96 12000 0.1771 24.8226 0.7654 0.7722 0.7681 0.7948 0.8019 0.7976 0.8016 0.8085 0.8043 0.9189 0.9265 0.9218 0.8667 0.875 0.8696 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 0.0909 0.0833
0.0596 1.12 14000 0.1725 23.7103 0.7773 0.7794 0.7777 0.8077 0.8100 0.8082 0.8148 0.8169 0.8151 0.9306 0.9334 0.9311 0.8889 0.8889 0.8889 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
0.0515 1.28 16000 0.1696 22.9305 0.7880 0.7897 0.7883 0.8183 0.8200 0.8185 0.8242 0.8261 0.8245 0.9352 0.9384 0.9360 0.9 0.9 0.8966 1.0 1.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.1111 0.1 0.1176
0.0369 1.44 18000 0.1695 22.4390 0.7937 0.7924 0.7925 0.8239 0.8226 0.8226 0.8297 0.8289 0.8286 0.9370 0.9391 0.9372 0.9 0.9 0.9 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
0.035 1.6 20000 0.1699 22.2358 0.7934 0.7948 0.7935 0.8227 0.8242 0.8228 0.8288 0.8303 0.8289 0.9380 0.9412 0.9388 0.9 0.9 0.9 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
0.0282 1.76 22000 0.1686 21.9549 0.7956 0.7940 0.7942 0.8258 0.8241 0.8243 0.8314 0.8298 0.8300 0.9408 0.9412 0.9402 0.9 0.9 0.9 1.0 1.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.1111 0.1 0.1176
0.0215 1.92 24000 0.1688 21.6445 0.8002 0.8022 0.8006 0.8287 0.8307 0.8291 0.8341 0.8363 0.8346 0.9380 0.9417 0.9391 0.9 0.9091 0.9 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
0.024 2.08 26000 0.1699 21.1899 0.8037 0.8070 0.8047 0.8331 0.8365 0.8342 0.8389 0.8424 0.8400 0.9415 0.9468 0.9434 0.9091 0.9091 0.9032 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
0.0198 2.24 28000 0.1696 21.3119 0.8038 0.8070 0.8048 0.8327 0.8360 0.8337 0.8382 0.8418 0.8393 0.9390 0.9445 0.9409 0.9091 0.9091 0.9032 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
0.0219 2.4 30000 0.1692 21.1641 0.8062 0.8080 0.8065 0.8352 0.8370 0.8355 0.8409 0.8430 0.8413 0.9407 0.9446 0.9419 0.9091 0.9091 0.9091 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

Framework versions

  • Transformers 4.39.0.dev0
  • Pytorch 2.2.1+cu121
  • Datasets 2.16.1
  • Tokenizers 0.15.0