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: 0.40436835891381345
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: 0.5951
- Wer Ortho: 0.4781
- Wer: 0.4044
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-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 |
|---|---|---|---|---|---|
| 2.605 | 1.79 | 50 | 2.3450 | 0.5355 | 0.3967 |
| 1.67 | 3.57 | 100 | 1.4800 | 0.5355 | 0.4126 |
| 0.8205 | 5.36 | 150 | 0.8745 | 0.5836 | 0.4787 |
| 0.5984 | 7.14 | 200 | 0.7396 | 0.4923 | 0.4079 |
| 0.4993 | 8.93 | 250 | 0.6831 | 0.4769 | 0.3996 |
| 0.4134 | 10.71 | 300 | 0.6510 | 0.4830 | 0.4032 |
| 0.384 | 12.5 | 350 | 0.6307 | 0.4738 | 0.3961 |
| 0.3286 | 14.29 | 400 | 0.6162 | 0.4806 | 0.4050 |
| 0.3188 | 16.07 | 450 | 0.6062 | 0.4800 | 0.4050 |
| 0.2751 | 17.86 | 500 | 0.6010 | 0.4843 | 0.4097 |
| 0.2568 | 19.64 | 550 | 0.5970 | 0.4750 | 0.4026 |
| 0.237 | 21.43 | 600 | 0.5951 | 0.4781 | 0.4044 |
Framework versions
- Transformers 4.38.0.dev0
- Pytorch 2.1.2
- Datasets 2.16.1
- Tokenizers 0.15.0