metadata
library_name: transformers
license: apache-2.0
base_model: openai/whisper-base
tags:
- generated_from_trainer
metrics:
- wer
model-index:
- name: whisper-base-en
results: []
whisper-base-en
This model is a fine-tuned version of openai/whisper-base on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.1212
- Wer: 3.6561
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
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- training_steps: 4000
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 0.1587 | 0.3676 | 100 | 0.1626 | 7.1105 |
| 0.1464 | 0.7353 | 200 | 0.1325 | 5.7486 |
| 0.0699 | 1.1029 | 300 | 0.1217 | 4.3894 |
| 0.0714 | 1.4706 | 400 | 0.1147 | 4.2034 |
| 0.0529 | 1.8382 | 500 | 0.1117 | 4.0358 |
| 0.0315 | 2.2059 | 600 | 0.1087 | 3.8865 |
| 0.0305 | 2.5735 | 700 | 0.1077 | 3.8787 |
| 0.0307 | 2.9412 | 800 | 0.1031 | 3.5958 |
| 0.0137 | 3.3088 | 900 | 0.1075 | 3.5304 |
| 0.0125 | 3.6765 | 1000 | 0.1065 | 3.4858 |
| 0.0103 | 4.0441 | 1100 | 0.1069 | 3.5592 |
| 0.0066 | 4.4118 | 1200 | 0.1093 | 3.5539 |
| 0.0063 | 4.7794 | 1300 | 0.1072 | 4.0332 |
| 0.0043 | 5.1471 | 1400 | 0.1095 | 3.5880 |
| 0.0045 | 5.5147 | 1500 | 0.1109 | 5.1672 |
| 0.0048 | 5.8824 | 1600 | 0.1114 | 3.5723 |
| 0.0035 | 6.25 | 1700 | 0.1128 | 3.5775 |
| 0.0033 | 6.6176 | 1800 | 0.1117 | 4.6591 |
| 0.0032 | 6.9853 | 1900 | 0.1132 | 3.5435 |
| 0.0032 | 7.3529 | 2000 | 0.1138 | 3.5801 |
| 0.0026 | 7.7206 | 2100 | 0.1151 | 3.6246 |
| 0.0024 | 8.0882 | 2200 | 0.1155 | 3.6639 |
| 0.0023 | 8.4559 | 2300 | 0.1167 | 3.6613 |
| 0.0022 | 8.8235 | 2400 | 0.1176 | 3.6299 |
| 0.0019 | 9.1912 | 2500 | 0.1177 | 3.5592 |
| 0.0018 | 9.5588 | 2600 | 0.1169 | 3.5827 |
| 0.0018 | 9.9265 | 2700 | 0.1175 | 3.5985 |
| 0.0016 | 10.2941 | 2800 | 0.1183 | 3.6142 |
| 0.0017 | 10.6618 | 2900 | 0.1190 | 3.6246 |
| 0.0016 | 11.0294 | 3000 | 0.1184 | 3.6954 |
| 0.0016 | 11.3971 | 3100 | 0.1192 | 3.6194 |
| 0.0015 | 11.7647 | 3200 | 0.1197 | 3.6508 |
| 0.0014 | 12.1324 | 3300 | 0.1202 | 3.6142 |
| 0.0013 | 12.5 | 3400 | 0.1202 | 3.6194 |
| 0.0014 | 12.8676 | 3500 | 0.1204 | 3.6561 |
| 0.0013 | 13.2353 | 3600 | 0.1208 | 3.6351 |
| 0.0014 | 13.6029 | 3700 | 0.1209 | 3.6561 |
| 0.0013 | 13.9706 | 3800 | 0.1211 | 3.6456 |
| 0.0014 | 14.3382 | 3900 | 0.1212 | 3.6613 |
| 0.0013 | 14.7059 | 4000 | 0.1212 | 3.6561 |
Framework versions
- Transformers 4.44.2
- Pytorch 2.5.0+cu121
- Datasets 3.1.0
- Tokenizers 0.19.1