|
--- |
|
library_name: transformers |
|
language: |
|
- ko |
|
license: apache-2.0 |
|
base_model: openai/whisper-small |
|
tags: |
|
- whisper-event |
|
- generated_from_trainer |
|
datasets: |
|
- GGarri/241113_newdata |
|
metrics: |
|
- wer |
|
model-index: |
|
- name: Whisper Small ko |
|
results: |
|
- task: |
|
name: Automatic Speech Recognition |
|
type: automatic-speech-recognition |
|
dataset: |
|
name: customdata |
|
type: GGarri/241113_newdata |
|
metrics: |
|
- name: Wer |
|
type: wer |
|
value: 0.8156606851549755 |
|
--- |
|
|
|
<!-- 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 ko |
|
|
|
This model is a fine-tuned version of [openai/whisper-small](https://huggingface.co/openai/whisper-small) on the customdata dataset. |
|
It achieves the following results on the evaluation set: |
|
- Loss: 0.0498 |
|
- Cer: 1.1070 |
|
- Wer: 0.8157 |
|
|
|
## 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: 32 |
|
- eval_batch_size: 8 |
|
- seed: 42 |
|
- optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments |
|
- lr_scheduler_type: linear |
|
- lr_scheduler_warmup_steps: 500 |
|
- training_steps: 5000 |
|
- mixed_precision_training: Native AMP |
|
|
|
### Training results |
|
|
|
| Training Loss | Epoch | Step | Validation Loss | Cer | Wer | |
|
|:-------------:|:-------:|:----:|:---------------:|:-------:|:-------:| |
|
| 1.1429 | 1.5625 | 100 | 0.8829 | 14.7984 | 14.5304 | |
|
| 0.3401 | 3.125 | 200 | 0.2637 | 2.0625 | 1.7828 | |
|
| 0.0413 | 4.6875 | 300 | 0.0599 | 1.5498 | 1.3167 | |
|
| 0.0163 | 6.25 | 400 | 0.0462 | 1.2818 | 0.9904 | |
|
| 0.0127 | 7.8125 | 500 | 0.0517 | 1.5265 | 1.1885 | |
|
| 0.0065 | 9.375 | 600 | 0.0402 | 1.5031 | 1.0487 | |
|
| 0.0028 | 10.9375 | 700 | 0.0396 | 1.7012 | 1.3167 | |
|
| 0.001 | 12.5 | 800 | 0.0406 | 1.5148 | 1.1186 | |
|
| 0.0004 | 14.0625 | 900 | 0.0405 | 1.4216 | 1.0371 | |
|
| 0.0005 | 15.625 | 1000 | 0.0424 | 1.5847 | 1.1885 | |
|
| 0.0001 | 17.1875 | 1100 | 0.0425 | 1.2701 | 0.9788 | |
|
| 0.0001 | 18.75 | 1200 | 0.0429 | 1.3051 | 1.0137 | |
|
| 0.0001 | 20.3125 | 1300 | 0.0432 | 1.2701 | 0.9788 | |
|
| 0.0001 | 21.875 | 1400 | 0.0436 | 1.2818 | 0.9904 | |
|
| 0.0001 | 23.4375 | 1500 | 0.0439 | 1.2934 | 1.0021 | |
|
| 0.0001 | 25.0 | 1600 | 0.0441 | 1.2934 | 1.0021 | |
|
| 0.0001 | 26.5625 | 1700 | 0.0443 | 1.2934 | 1.0021 | |
|
| 0.0001 | 28.125 | 1800 | 0.0446 | 1.2934 | 1.0021 | |
|
| 0.0001 | 29.6875 | 1900 | 0.0448 | 1.2818 | 0.9904 | |
|
| 0.0001 | 31.25 | 2000 | 0.0449 | 1.2002 | 0.9089 | |
|
| 0.0001 | 32.8125 | 2100 | 0.0454 | 1.2002 | 0.9089 | |
|
| 0.0001 | 34.375 | 2200 | 0.0458 | 1.2002 | 0.9089 | |
|
| 0.0 | 35.9375 | 2300 | 0.0461 | 1.2002 | 0.9089 | |
|
| 0.0 | 37.5 | 2400 | 0.0463 | 1.1769 | 0.8856 | |
|
| 0.0 | 39.0625 | 2500 | 0.0465 | 1.1769 | 0.8856 | |
|
| 0.0 | 40.625 | 2600 | 0.0467 | 1.1536 | 0.8623 | |
|
| 0.0 | 42.1875 | 2700 | 0.0469 | 1.1303 | 0.8390 | |
|
| 0.0 | 43.75 | 2800 | 0.0471 | 1.1536 | 0.8623 | |
|
| 0.0 | 45.3125 | 2900 | 0.0473 | 1.1536 | 0.8623 | |
|
| 0.0 | 46.875 | 3000 | 0.0474 | 1.1536 | 0.8623 | |
|
| 0.0 | 48.4375 | 3100 | 0.0476 | 1.1536 | 0.8623 | |
|
| 0.0 | 50.0 | 3200 | 0.0477 | 1.1303 | 0.8390 | |
|
| 0.0 | 51.5625 | 3300 | 0.0478 | 1.1419 | 0.8506 | |
|
| 0.0 | 53.125 | 3400 | 0.0479 | 1.1186 | 0.8273 | |
|
| 0.0 | 54.6875 | 3500 | 0.0481 | 1.1186 | 0.8273 | |
|
| 0.0 | 56.25 | 3600 | 0.0482 | 1.1186 | 0.8273 | |
|
| 0.0 | 57.8125 | 3700 | 0.0483 | 1.1186 | 0.8273 | |
|
| 0.0 | 59.375 | 3800 | 0.0484 | 1.1070 | 0.8157 | |
|
| 0.0 | 60.9375 | 3900 | 0.0485 | 1.1070 | 0.8157 | |
|
| 0.0 | 62.5 | 4000 | 0.0487 | 1.1070 | 0.8157 | |
|
| 0.0 | 64.0625 | 4100 | 0.0490 | 1.1070 | 0.8157 | |
|
| 0.0 | 65.625 | 4200 | 0.0492 | 1.1070 | 0.8157 | |
|
| 0.0 | 67.1875 | 4300 | 0.0494 | 1.1070 | 0.8157 | |
|
| 0.0 | 68.75 | 4400 | 0.0495 | 1.1070 | 0.8157 | |
|
| 0.0 | 70.3125 | 4500 | 0.0496 | 1.1070 | 0.8157 | |
|
| 0.0 | 71.875 | 4600 | 0.0497 | 1.1070 | 0.8157 | |
|
| 0.0 | 73.4375 | 4700 | 0.0497 | 1.1070 | 0.8157 | |
|
| 0.0 | 75.0 | 4800 | 0.0497 | 1.1070 | 0.8157 | |
|
| 0.0 | 76.5625 | 4900 | 0.0498 | 1.1070 | 0.8157 | |
|
| 0.0 | 78.125 | 5000 | 0.0498 | 1.1070 | 0.8157 | |
|
|
|
|
|
### Framework versions |
|
|
|
- Transformers 4.46.2 |
|
- Pytorch 2.4.0 |
|
- Datasets 2.18.0 |
|
- Tokenizers 0.20.3 |
|
|
