|
--- |
|
language: |
|
- fa |
|
license: apache-2.0 |
|
base_model: openai/whisper-small |
|
tags: |
|
- fa-asr |
|
- generated_from_trainer |
|
datasets: |
|
- mozilla-foundation/common_voice_15_0 |
|
metrics: |
|
- wer |
|
model-index: |
|
- name: Whisper Small Persian |
|
results: [] |
|
--- |
|
|
|
<!-- This model card has been generated automatically according to the information the Trainer had access to. You |
|
should probably proofread and complete it, then remove this comment. --> |
|
|
|
# Whisper Small Persian |
|
|
|
This model is a fine-tuned version of [openai/whisper-small](https://huggingface.co/openai/whisper-small) on the Common Voice 15.0 dataset. |
|
It achieves the following results on the evaluation set: |
|
- Loss: 0.9061 |
|
- Wer: 51.3701 |
|
|
|
## 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: 14 |
|
- eval_batch_size: 10 |
|
- seed: 42 |
|
- gradient_accumulation_steps: 4 |
|
- total_train_batch_size: 56 |
|
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 |
|
- lr_scheduler_type: linear |
|
- lr_scheduler_warmup_steps: 50 |
|
- training_steps: 10000 |
|
- mixed_precision_training: Native AMP |
|
|
|
### Training results |
|
|
|
| Training Loss | Epoch | Step | Validation Loss | Wer | |
|
|:-------------:|:-----:|:-----:|:---------------:|:-------:| |
|
| 0.7495 | 0.39 | 100 | 0.8602 | 58.0774 | |
|
| 0.3059 | 0.77 | 200 | 0.5538 | 55.2248 | |
|
| 0.2002 | 1.16 | 300 | 0.5137 | 52.4615 | |
|
| 0.1798 | 1.54 | 400 | 0.5103 | 51.6578 | |
|
| 0.1678 | 1.93 | 500 | 0.4806 | 46.5512 | |
|
| 0.0932 | 2.32 | 600 | 0.4936 | 47.4955 | |
|
| 0.0944 | 2.7 | 700 | 0.5120 | 48.9160 | |
|
| 0.0493 | 3.09 | 800 | 0.5198 | 48.2611 | |
|
| 0.0506 | 3.47 | 900 | 0.5227 | 47.4128 | |
|
| 0.0482 | 3.86 | 1000 | 0.5256 | 46.9498 | |
|
| 0.0244 | 4.25 | 1100 | 0.5491 | 46.0849 | |
|
| 0.0272 | 4.63 | 1200 | 0.5621 | 46.9167 | |
|
| 0.0246 | 5.02 | 1300 | 0.5856 | 48.8036 | |
|
| 0.0159 | 5.41 | 1400 | 0.6145 | 48.3041 | |
|
| 0.0152 | 5.79 | 1500 | 0.6178 | 47.7568 | |
|
| 0.008 | 6.18 | 1600 | 0.6191 | 48.0611 | |
|
| 0.0077 | 6.56 | 1700 | 0.6309 | 46.3594 | |
|
| 0.0097 | 6.95 | 1800 | 0.6272 | 47.9139 | |
|
| 0.0056 | 7.34 | 1900 | 0.6594 | 46.5562 | |
|
| 0.0049 | 7.72 | 2000 | 0.6581 | 47.5831 | |
|
| 0.0042 | 8.11 | 2100 | 0.6953 | 48.0925 | |
|
| 0.0038 | 8.49 | 2200 | 0.6857 | 48.3719 | |
|
| 0.0033 | 8.88 | 2300 | 0.6983 | 49.6965 | |
|
| 0.003 | 9.27 | 2400 | 0.7109 | 48.1173 | |
|
| 0.0033 | 9.65 | 2500 | 0.6899 | 48.2363 | |
|
| 0.0027 | 10.04 | 2600 | 0.7074 | 48.3405 | |
|
| 0.0035 | 10.42 | 2700 | 0.7018 | 47.1003 | |
|
| 0.0026 | 10.81 | 2800 | 0.7198 | 47.1251 | |
|
| 0.0027 | 11.2 | 2900 | 0.7460 | 48.2760 | |
|
| 0.0023 | 11.58 | 3000 | 0.7348 | 47.7154 | |
|
| 0.0023 | 11.97 | 3100 | 0.7387 | 48.0081 | |
|
| 0.0024 | 12.36 | 3200 | 0.7199 | 46.8687 | |
|
| 0.0033 | 12.74 | 3300 | 0.7250 | 48.6895 | |
|
| 0.0017 | 13.13 | 3400 | 0.7242 | 49.0450 | |
|
| 0.0016 | 13.51 | 3500 | 0.7359 | 48.7010 | |
|
| 0.0022 | 13.9 | 3600 | 0.7220 | 48.1371 | |
|
| 0.0016 | 14.29 | 3700 | 0.7431 | 46.7563 | |
|
| 0.0012 | 14.67 | 3800 | 0.7564 | 47.0143 | |
|
| 0.0014 | 15.06 | 3900 | 0.7770 | 47.0242 | |
|
| 0.0008 | 15.44 | 4000 | 0.8116 | 48.4695 | |
|
| 0.0015 | 15.83 | 4100 | 0.7623 | 48.0925 | |
|
| 0.002 | 16.22 | 4200 | 0.7699 | 47.7254 | |
|
| 0.001 | 16.6 | 4300 | 0.7631 | 46.8754 | |
|
| 0.0009 | 16.99 | 4400 | 0.7591 | 48.4166 | |
|
| 0.0008 | 17.37 | 4500 | 0.7797 | 47.1780 | |
|
| 0.0008 | 17.76 | 4600 | 0.7851 | 46.7662 | |
|
| 0.0007 | 18.15 | 4700 | 0.7897 | 48.3157 | |
|
| 0.0006 | 18.53 | 4800 | 0.7760 | 48.8069 | |
|
| 0.0016 | 18.92 | 4900 | 0.7763 | 47.9718 | |
|
| 0.0009 | 19.31 | 5000 | 0.8151 | 48.1586 | |
|
| 0.0004 | 19.69 | 5100 | 0.7967 | 47.5567 | |
|
| 0.0007 | 20.08 | 5200 | 0.8094 | 46.7861 | |
|
| 0.001 | 20.46 | 5300 | 0.8206 | 47.1565 | |
|
| 0.0008 | 20.85 | 5400 | 0.8015 | 47.6212 | |
|
| 0.0004 | 21.24 | 5500 | 0.8104 | 47.2210 | |
|
| 0.0003 | 21.62 | 5600 | 0.8020 | 48.1454 | |
|
| 0.0004 | 22.01 | 5700 | 0.8295 | 46.9415 | |
|
| 0.0004 | 22.39 | 5800 | 0.8228 | 46.6108 | |
|
| 0.0005 | 22.78 | 5900 | 0.8386 | 48.0974 | |
|
| 0.0007 | 23.17 | 6000 | 0.8415 | 46.6141 | |
|
| 0.0003 | 23.55 | 6100 | 0.8283 | 46.3263 | |
|
| 0.0005 | 23.94 | 6200 | 0.8342 | 45.9658 | |
|
| 0.0002 | 24.32 | 6300 | 0.8379 | 46.6240 | |
|
| 0.0005 | 24.71 | 6400 | 0.8371 | 48.0809 | |
|
| 0.0002 | 25.1 | 6500 | 0.8258 | 47.5864 | |
|
| 0.0001 | 25.48 | 6600 | 0.8396 | 47.2607 | |
|
| 0.0002 | 25.87 | 6700 | 0.8491 | 47.5435 | |
|
| 0.0005 | 26.25 | 6800 | 0.8444 | 48.9937 | |
|
| 0.0001 | 26.64 | 6900 | 0.8540 | 48.0544 | |
|
| 0.0001 | 27.03 | 7000 | 0.8605 | 48.5390 | |
|
| 0.0001 | 27.41 | 7100 | 0.8613 | 48.9226 | |
|
| 0.0001 | 27.8 | 7200 | 0.8673 | 48.9491 | |
|
| 0.0001 | 28.19 | 7300 | 0.8688 | 48.7936 | |
|
| 0.0001 | 28.57 | 7400 | 0.8711 | 48.8350 | |
|
| 0.0001 | 28.96 | 7500 | 0.8728 | 48.8581 | |
|
| 0.0001 | 29.34 | 7600 | 0.8736 | 49.8454 | |
|
| 0.0001 | 29.73 | 7700 | 0.8759 | 49.9248 | |
|
| 0.0001 | 30.12 | 7800 | 0.8786 | 49.6932 | |
|
| 0.0001 | 30.5 | 7900 | 0.8809 | 49.7429 | |
|
| 0.0001 | 30.89 | 8000 | 0.8826 | 49.9132 | |
|
| 0.0001 | 31.27 | 8100 | 0.8840 | 50.4506 | |
|
| 0.0001 | 31.66 | 8200 | 0.8859 | 50.2919 | |
|
| 0.0001 | 32.05 | 8300 | 0.8888 | 50.5532 | |
|
| 0.0001 | 32.43 | 8400 | 0.8924 | 50.6491 | |
|
| 0.0001 | 32.82 | 8500 | 0.8928 | 50.9997 | |
|
| 0.0001 | 33.2 | 8600 | 0.8920 | 51.3188 | |
|
| 0.0001 | 33.59 | 8700 | 0.8946 | 51.1419 | |
|
| 0.0001 | 33.98 | 8800 | 0.8956 | 51.1551 | |
|
| 0.0001 | 34.36 | 8900 | 0.8988 | 50.7483 | |
|
| 0.0001 | 34.75 | 9000 | 0.9003 | 50.8889 | |
|
| 0.0001 | 35.14 | 9100 | 0.9011 | 50.9302 | |
|
| 0.0001 | 35.52 | 9200 | 0.9020 | 50.9533 | |
|
| 0.0001 | 35.91 | 9300 | 0.9019 | 51.5603 | |
|
| 0.0001 | 36.29 | 9400 | 0.9033 | 51.6479 | |
|
| 0.0001 | 36.68 | 9500 | 0.9040 | 51.6810 | |
|
| 0.0001 | 37.07 | 9600 | 0.9049 | 51.6744 | |
|
| 0.0001 | 37.45 | 9700 | 0.9053 | 51.6429 | |
|
| 0.0001 | 37.84 | 9800 | 0.9057 | 51.5966 | |
|
| 0.0001 | 38.22 | 9900 | 0.9060 | 51.7372 | |
|
| 0.0001 | 38.61 | 10000 | 0.9061 | 51.3701 | |
|
|
|
|
|
### Framework versions |
|
|
|
- Transformers 4.35.0 |
|
- Pytorch 2.0.1+cu117 |
|
- Datasets 2.14.6 |
|
- Tokenizers 0.14.1 |
|
|
