metadata
license: apache-2.0
base_model: openai/whisper-tiny
tags:
- generated_from_trainer
datasets:
- PolyAI/minds14
metrics:
- wer
model-index:
- name: whisper-tiny-finetuned-minds14
results:
- task:
name: Automatic Speech Recognition
type: automatic-speech-recognition
dataset:
name: PolyAI/minds14
type: PolyAI/minds14
config: en-US
split: train
args: en-US
metrics:
- name: Wer
type: wer
value: 0.3624031007751938
whisper-tiny-finetuned-minds14
This model is a fine-tuned version of openai/whisper-tiny on the PolyAI/minds14 dataset. It achieves the following results on the evaluation set:
- Loss: 0.6785
- Wer Ortho: 0.3607
- Wer: 0.3624
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: 16
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: constant_with_warmup
- lr_scheduler_warmup_ratio: 0.1
- lr_scheduler_warmup_steps: 50
- training_steps: 500
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer Ortho | Wer |
|---|---|---|---|---|---|
| 3.8342 | 1.0 | 28 | 2.7013 | 0.4859 | 0.3669 |
| 1.52 | 2.0 | 56 | 0.6447 | 0.3822 | 0.3624 |
| 0.4282 | 3.0 | 84 | 0.5154 | 0.3573 | 0.3521 |
| 0.2511 | 4.0 | 112 | 0.5017 | 0.3452 | 0.3430 |
| 0.1461 | 5.0 | 140 | 0.5106 | 0.3620 | 0.3572 |
| 0.0829 | 6.0 | 168 | 0.5399 | 0.3641 | 0.3592 |
| 0.0423 | 7.0 | 196 | 0.5596 | 0.3573 | 0.3527 |
| 0.0199 | 8.0 | 224 | 0.5846 | 0.3627 | 0.3598 |
| 0.0093 | 9.0 | 252 | 0.6006 | 0.3594 | 0.3572 |
| 0.0056 | 10.0 | 280 | 0.6207 | 0.3345 | 0.3301 |
| 0.0037 | 11.0 | 308 | 0.6238 | 0.3560 | 0.3534 |
| 0.0021 | 12.0 | 336 | 0.6377 | 0.3486 | 0.3482 |
| 0.0016 | 13.0 | 364 | 0.6485 | 0.3594 | 0.3579 |
| 0.0013 | 14.0 | 392 | 0.6621 | 0.3567 | 0.3572 |
| 0.0011 | 15.0 | 420 | 0.6617 | 0.3587 | 0.3605 |
| 0.0009 | 16.0 | 448 | 0.6682 | 0.3560 | 0.3559 |
| 0.0008 | 17.0 | 476 | 0.6741 | 0.3627 | 0.3624 |
| 0.0008 | 17.86 | 500 | 0.6785 | 0.3607 | 0.3624 |
Framework versions
- Transformers 4.37.2
- Pytorch 2.1.2
- Datasets 2.1.0
- Tokenizers 0.15.1