metadata
license: apache-2.0
base_model: openai/whisper-tiny
tags:
- generated_from_trainer
datasets:
- PolyAI/minds14
metrics:
- wer
model-index:
- name: whisper-tiny-ft-PolyAI-minds-14-enUS
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: 0.39315230224321135
whisper-tiny-ft-PolyAI-minds-14-enUS
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: 0.7067
- Wer Ortho: 0.3991
- Wer: 0.3932
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: 4e-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_steps: 150
- training_steps: 300
Training results
Training Loss | Epoch | Step | Validation Loss | Wer Ortho | Wer |
---|---|---|---|---|---|
3.1391 | 0.89 | 25 | 2.1392 | 0.5392 | 0.4061 |
1.1631 | 1.79 | 50 | 0.6561 | 0.5811 | 0.5531 |
0.4253 | 2.68 | 75 | 0.5337 | 0.6397 | 0.6257 |
0.2072 | 3.57 | 100 | 0.5282 | 0.4756 | 0.4628 |
0.1265 | 4.46 | 125 | 0.5668 | 0.4300 | 0.4227 |
0.0507 | 5.36 | 150 | 0.5849 | 0.4485 | 0.4380 |
0.0357 | 6.25 | 175 | 0.6048 | 0.3584 | 0.3430 |
0.0243 | 7.14 | 200 | 0.6141 | 0.3590 | 0.3465 |
0.0154 | 8.04 | 225 | 0.6424 | 0.3701 | 0.3524 |
0.0189 | 8.93 | 250 | 0.6685 | 0.3732 | 0.3660 |
0.0154 | 9.82 | 275 | 0.6933 | 0.3924 | 0.3861 |
0.0091 | 10.71 | 300 | 0.7067 | 0.3991 | 0.3932 |
Framework versions
- Transformers 4.33.0
- Pytorch 1.12.1+cu116
- Datasets 2.14.4
- Tokenizers 0.12.1