openai/whisper-medium
This model is a fine-tuned version of openai/whisper-medium on the common_voice_11_0 dataset. It achieves the following results on the evaluation set:
- Loss: 0.3029
- Wer: 9.0355
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-05
- train_batch_size: 32
- eval_batch_size: 16
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 500
- training_steps: 5000
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 0.0392 | 3.03 | 1000 | 0.2023 | 10.1807 |
| 0.0036 | 7.01 | 2000 | 0.2478 | 9.4409 |
| 0.0013 | 10.04 | 3000 | 0.2791 | 9.1014 |
| 0.0002 | 14.01 | 4000 | 0.2970 | 9.0625 |
| 0.0002 | 17.04 | 5000 | 0.3029 | 9.0355 |
Framework versions
- Transformers 4.26.0.dev0
- Pytorch 1.13.0+cu117
- Datasets 2.7.1.dev0
- Tokenizers 0.13.2
- Downloads last month
- 5,115
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social
visibility and check back later, or deploy to Inference Endpoints (dedicated)
instead.
Dataset used to train vumichien/whisper-medium-jp
Spaces using vumichien/whisper-medium-jp 17
Evaluation results
- WER on mozilla-foundation/common_voice_11_0 jatest set self-reported9.035
- CER on mozilla-foundation/common_voice_11_0 jatest set self-reported5.610
- WER on google/fleurs ja_jptest set self-reported13.560
- CER on google/fleurs ja_jptest set self-reported8.010