whisper-tiny / README.md
bagasshw's picture
End of training
9ff1973 verified
metadata
library_name: transformers
language:
  - jw
license: apache-2.0
base_model: openai/whisper-tiny
tags:
  - generated_from_trainer
datasets:
  - SEACrowd/jv_id_tts
metrics:
  - wer
model-index:
  - name: Whisper Tiny Java - HQ TTS
    results:
      - task:
          name: Automatic Speech Recognition
          type: automatic-speech-recognition
        dataset:
          name: jv_id_tts
          type: SEACrowd/jv_id_tts
          config: jv_id_tts_source
          split: None
          args: 'config: jw, split: test'
        metrics:
          - name: Wer
            type: wer
            value: 20.152640264026402

Whisper Tiny Java - HQ TTS

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

  • Loss: 0.3168
  • Wer: 20.1526

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: 3.75e-05
  • train_batch_size: 16
  • eval_batch_size: 8
  • 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: 100
  • training_steps: 2500
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Validation Loss Wer
0.3173 1.7123 500 0.4364 30.1774
0.0564 3.4247 1000 0.3388 22.7929
0.0202 5.1370 1500 0.3240 20.8746
0.0055 6.8493 2000 0.3174 20.4620
0.003 8.5616 2500 0.3168 20.1526

Framework versions

  • Transformers 4.51.3
  • Pytorch 2.7.0+cu126
  • Datasets 2.18.0
  • Tokenizers 0.21.1