Hugging Face's logo

Baselhany
/
Whisper_tiny_fine_tune_Quran

Automatic Speech Recognition
Transformers
TensorBoard
Safetensors
Arabic
whisper
Generated from Trainer
Inference Endpoints
Model card Files Metrics Community
Whisper_tiny_fine_tune_Quran / README.md
Baselhany's picture
Baselhany
Training finished
fbffb6a verified
preview code
|
raw
history blame
5.74 kB
metadata
library_name: transformers
language:
  - ar
license: apache-2.0
base_model: openai/whisper-tiny
tags:
  - generated_from_trainer
metrics:
  - wer
model-index:
  - name: Whisper tiny AR - BH
    results: []

Whisper tiny AR - BH

This model is a fine-tuned version of openai/whisper-tiny on the quran-ayat-speech-to-text dataset. It achieves the following results on the evaluation set:

  • Loss: 0.0141
  • Wer: 12.6647
  • Cer: 4.0046

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: 0.0001
  • train_batch_size: 16
  • eval_batch_size: 16
  • seed: 42
  • gradient_accumulation_steps: 4
  • total_train_batch_size: 64
  • optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
  • lr_scheduler_type: cosine
  • lr_scheduler_warmup_steps: 500
  • num_epochs: 8
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Cer Validation Loss Wer
0.0258 0.1408 400 52.2218 0.0246 104.9348
0.0177 0.2817 800 10.2633 0.0184 26.2089
0.0116 0.4225 1200 7.3210 0.0160 20.9517
0.0101 0.5633 1600 5.8227 0.0141 17.5020
0.008 0.7042 2000 5.1235 0.0127 16.3695
0.0057 0.8450 2400 4.8168 0.0119 15.2343
0.0056 0.9858 2800 4.6678 0.0116 14.6364
0.0071 1.1267 3200 5.3042 0.0135 15.8929
0.0059 1.2676 3600 5.0437 0.0132 15.7165
0.0056 1.4084 4000 5.3648 0.0124 14.5758
0.0041 1.5492 4400 4.7531 0.0122 14.2259
0.0038 1.6901 4800 4.7431 0.0120 13.8043
0.004 1.8309 5200 4.9569 0.0119 14.1818
0.0036 1.9717 5600 4.9171 0.0118 14.0743
0.0033 2.1127 6000 5.0453 0.0129 15.0828
0.0033 2.2535 6400 5.1424 0.0128 14.9340
0.0033 2.3943 6800 5.0171 0.0123 14.7329
0.0033 2.5352 7200 4.3676 0.0122 13.6748
0.0034 2.6760 7600 4.5300 0.0122 13.5618
0.0025 2.8168 8000 4.4698 0.0122 13.3662
0.0028 2.9577 8400 4.5794 0.0122 13.5536
0.003 3.0986 8800 5.0764 0.0125 15.1021
0.0024 3.2394 9200 5.1331 0.0125 14.6943
0.0019 3.3802 9600 5.8448 0.0128 16.2924
0.0023 3.5211 10000 5.1642 0.0128 14.7301
0.002 3.6619 10400 4.9046 0.0127 13.8649
0.0018 3.8027 10800 4.9748 0.0126 13.6610
0.0021 3.9436 11200 5.0136 0.0126 13.8539
0.0018 4.0845 11600 5.0283 0.0132 14.6475
0.0018 4.2253 12000 4.5932 0.0132 13.7988
0.0022 4.3662 12400 4.3948 0.0130 13.7354
0.0025 4.5070 12800 4.7691 0.0131 14.3774
0.0018 4.6478 13200 4.8726 0.0131 14.0854
0.0016 4.7887 13600 4.7136 0.0130 14.0165
0.0018 4.9295 14000 4.7886 0.0130 14.0661
0.0017 5.0704 14400 4.5393 0.0133 14.0110
0.0013 5.2112 14800 4.3028 0.0132 13.7547
0.0017 5.3521 15200 4.5275 0.0133 14.2231
0.0014 5.4929 15600 4.6271 0.0135 14.1983
0.0016 5.6337 16000 4.3983 0.0134 13.8539
0.0015 5.7746 16400 4.2035 0.0134 13.5426
0.0016 5.9154 16800 4.2561 0.0134 13.6335
0.0015 6.0563 17200 4.3246 0.0134 13.6059
0.0015 6.1972 17600 4.1759 0.0137 13.6142
0.0016 6.3380 18000 4.2195 0.0137 13.5536
0.0014 6.4788 18400 4.4176 0.0137 13.8760
0.0015 6.6197 18800 4.2144 0.0137 13.5784
0.0015 6.7605 19200 4.1868 0.0137 13.4874
0.0016 6.9013 19600 4.0946 0.0137 13.3442
0.0015 7.0422 20000 0.0139 13.5508 4.1526
0.0012 7.1831 20400 0.0139 13.5040 4.1830
0.0011 7.3239 20800 0.0138 13.3194 4.0708
0.0017 7.4647 21200 0.0138 13.3552 4.0446
0.0012 7.6056 21600 0.0139 13.3194 4.0699
0.0011 7.7464 22000 0.0140 13.3001 4.0378
0.0012 7.8872 22400 0.0139 13.3442 4.0558

Framework versions

  • Transformers 4.47.0
  • Pytorch 2.5.1+cu121
  • Datasets 3.2.0
  • Tokenizers 0.21.0