Whisper tiny Japanese

This model is a fine-tuned version of openai/whisper-tiny on the Japanese English dataset. It achieves the following results on the evaluation set:

  • Loss: 0.4624
  • Wer: 21.5069

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: 2
  • eval_batch_size: 1
  • seed: 42
  • optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
  • lr_scheduler_type: linear
  • lr_scheduler_warmup_steps: 500
  • training_steps: 2000

Training results

Training Loss Epoch Step Validation Loss Wer
0.446 0.2488 200 0.6420 30.8741
0.3568 0.4975 400 0.5394 26.0182
0.2883 0.7463 600 0.5164 23.2926
0.41 0.9950 800 0.4806 23.1673
0.137 1.2438 1000 0.4815 22.1648
0.167 1.4925 1200 0.4720 21.6949
0.0589 1.7413 1400 0.4677 21.8672
0.1126 1.9900 1600 0.4565 21.9142
0.0299 2.2388 1800 0.4622 21.5382
0.0899 2.4876 2000 0.4624 21.5069

Framework versions

  • Transformers 4.50.0.dev0
  • Pytorch 2.5.1+cu124
  • Datasets 3.3.2
  • Tokenizers 0.21.0
Downloads last month
8
Safetensors
Model size
37.8M params
Tensor type
F32
·
Inference Providers NEW
This model is not currently available via any of the supported Inference Providers.

Model tree for DeathCover1003/whisper-tiny_to_japanese_accent_2000_200

Finetuned
(1386)
this model

Evaluation results