metadata
license: apache-2.0
base_model: openai/whisper-tiny
tags:
- whisper-event
- generated_from_trainer
datasets:
- common_voice_11_0
metrics:
- wer
model-index:
- name: WhisperTinyFinnishV3
results:
- task:
name: Automatic Speech Recognition
type: automatic-speech-recognition
dataset:
name: common_voice_11_0
type: common_voice_11_0
config: fi
split: test
args: fi
metrics:
- name: Wer
type: wer
value: 45.13758009800226
WhisperTinyFinnishV3
This model is a fine-tuned version of openai/whisper-tiny on the common_voice_11_0 dataset. It achieves the following results on the evaluation set:
- Loss: 0.5363
- Wer: 45.1376
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: 3e-06
- train_batch_size: 32
- eval_batch_size: 4
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 1000
- training_steps: 10000
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 0.9236 | 0.1 | 1000 | 0.7783 | 58.5187 |
| 0.727 | 0.2 | 2000 | 0.6638 | 53.1097 |
| 0.6867 | 0.3 | 3000 | 0.6113 | 50.2639 |
| 0.8348 | 0.4 | 4000 | 0.5882 | 48.2661 |
| 0.5165 | 0.5 | 5000 | 0.5679 | 47.1259 |
| 0.5509 | 0.6 | 6000 | 0.5540 | 46.6359 |
| 0.639 | 0.7 | 7000 | 0.5466 | 46.5228 |
| 0.4715 | 0.8 | 8000 | 0.5400 | 45.9763 |
| 0.6306 | 0.9 | 9000 | 0.5363 | 45.1376 |
| 0.4598 | 1.0 | 10000 | 0.5352 | 45.4768 |
Framework versions
- Transformers 4.36.2
- Pytorch 2.1.0+cu121
- Datasets 2.16.1
- Tokenizers 0.15.0