metadata
license: apache-2.0
base_model: openai/whisper-tiny
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: PolyAI/minds14
type: PolyAI/minds14
config: en-US
split: train
args: en-US
metrics:
- name: Wer
type: wer
value: 0.6735537190082644
whisper-tiny
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: 0.7357
- Wer Ortho: 0.6860
- Wer: 0.6736
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: 5e-06
- 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: 50
- training_steps: 600
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer Ortho | Wer |
|---|---|---|---|---|---|
| 2.2141 | 1.79 | 50 | 1.4145 | 0.5077 | 0.4044 |
| 0.4779 | 3.57 | 100 | 0.6149 | 0.4571 | 0.3967 |
| 0.2486 | 5.36 | 150 | 0.5744 | 0.4405 | 0.3808 |
| 0.1641 | 7.14 | 200 | 0.5754 | 0.4368 | 0.3802 |
| 0.0912 | 8.93 | 250 | 0.5966 | 0.4399 | 0.3985 |
| 0.0385 | 10.71 | 300 | 0.6222 | 0.4324 | 0.3996 |
| 0.0203 | 12.5 | 350 | 0.6532 | 0.4861 | 0.4616 |
| 0.0079 | 14.29 | 400 | 0.6867 | 0.5503 | 0.5331 |
| 0.0039 | 16.07 | 450 | 0.7002 | 0.5713 | 0.5555 |
| 0.0026 | 17.86 | 500 | 0.7144 | 0.6428 | 0.6275 |
| 0.0021 | 19.64 | 550 | 0.7275 | 0.6619 | 0.6499 |
| 0.0017 | 21.43 | 600 | 0.7357 | 0.6860 | 0.6736 |
Framework versions
- Transformers 4.38.0.dev0
- Pytorch 2.1.2
- Datasets 2.16.1
- Tokenizers 0.15.0