Hugging Face's logo

vumichien
/
whisper-small-ja

Automatic Speech Recognition Transformers PyTorch TensorBoard Safetensors
Japanese whisper whisper-event generated_from_trainer Eval Results Inference Endpoints
Model card Files Metrics Community
2
whisper-small-ja / README.md
vumichien's picture
vumichien
Librarian Bot: Add base_model information to model (#2)
123b477
preview code
|
raw history blame contribute delete
No virus
2.58 kB
metadata
language:
  - ja
license: apache-2.0
tags:
  - whisper-event
  - generated_from_trainer
datasets:
  - mozilla-foundation/common_voice_11_0
metrics:
  - wer
base_model: openai/whisper-small
model-index:
  - name: Whisper Small Japanese
    results:
      - task:
          type: automatic-speech-recognition
          name: Automatic Speech Recognition
        dataset:
          name: mozilla-foundation/common_voice_11_0 ja
          type: mozilla-foundation/common_voice_11_0
          config: ja
          split: test
          args: ja
        metrics:
          - type: wer
            value: 13.467905405405405
            name: Wer
          - type: cer
            value: 8.6022
            name: Cer
      - task:
          type: automatic-speech-recognition
          name: Automatic Speech Recognition
        dataset:
          name: google/fleurs
          type: google/fleurs
          config: ja_jp
          split: test
        metrics:
          - type: wer
            value: 21.46
            name: WER
          - type: cer
            value: 13.65
            name: CER

Whisper Small Japanese

This model is a fine-tuned version of openai/whisper-small on the mozilla-foundation/common_voice_11_0 ja dataset. It achieves the following results on the evaluation set:

  • Loss: 0.4232
  • Wer: 13.4679
  • Cer: 8.6022

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: 64
  • eval_batch_size: 32
  • 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 Cer
0.0167 7.0 1000 0.3066 13.6740 8.5733
0.0021 14.01 2000 0.3579 13.8733 8.7816
0.0006 21.01 3000 0.4025 13.5794 8.6173
0.0004 28.01 4000 0.4232 13.4679 8.6022
0.0004 35.01 5000 0.4319 13.4747 8.6213

Framework versions

  • Transformers 4.26.0.dev0
  • Pytorch 1.12.1+cu113
  • Datasets 2.7.1.dev0
  • Tokenizers 0.13.2