metadata
license: apache-2.0
tags:
- generated_from_trainer
datasets:
- PolyAI/minds14
metrics:
- wer
model-index:
- name: Whisper Tiny
results:
- task:
name: Automatic Speech Recognition
type: automatic-speech-recognition
dataset:
name: Minds 14
type: PolyAI/minds14
config: en-US
split: train
args: en-US
metrics:
- name: Wer
type: wer
value: 0.333530106257379
Whisper Tiny
This model is a fine-tuned version of openai/whisper-tiny on the Minds 14 dataset. It achieves the following results on the evaluation set:
- Loss: 0.6262
- Wer Ortho: 0.3455
- Wer: 0.3335
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: 2e-05
- train_batch_size: 32
- eval_batch_size: 32
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 10
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer Ortho | Wer |
|---|---|---|---|---|---|
| 3.7423 | 1.0 | 15 | 2.9584 | 0.5275 | 0.4079 |
| 0.5112 | 2.0 | 30 | 0.7722 | 0.4022 | 0.3731 |
| 0.2542 | 3.0 | 45 | 0.6002 | 0.3837 | 0.3619 |
| 0.1196 | 4.0 | 60 | 0.5739 | 0.3492 | 0.3294 |
| 0.0214 | 5.0 | 75 | 0.5843 | 0.3652 | 0.3542 |
| 0.0659 | 6.0 | 90 | 0.6047 | 0.3418 | 0.3282 |
| 0.0322 | 7.0 | 105 | 0.6134 | 0.3560 | 0.3424 |
| 0.0049 | 8.0 | 120 | 0.6180 | 0.3553 | 0.3406 |
| 0.0348 | 9.0 | 135 | 0.6242 | 0.3442 | 0.3323 |
| 0.1332 | 10.0 | 150 | 0.6262 | 0.3455 | 0.3335 |
Framework versions
- Transformers 4.31.0.dev0
- Pytorch 1.13.1
- Datasets 2.13.1
- Tokenizers 0.13.2