csikasote's picture
Update metadata with huggingface_hub
d4bdb08 verified
metadata
base_model: openai/whisper-large-v3
datasets:
  - BembaSpeech
license: apache-2.0
metrics:
  - wer
tags:
  - generated_from_trainer
model-index:
  - name: whisper-large-v3-bem
    results:
      - task:
          type: automatic-speech-recognition
          name: Automatic Speech Recognition
        dataset:
          name: BembaSpeech bem
          type: BembaSpeech
          args: bem
        metrics:
          - type: wer
            value: 0.375750300120048
            name: Wer
      - task:
          type: automatic-speech-recognition
          name: Automatic Speech Recognition
        dataset:
          name: bembaspeech
          type: bembaspeech
          config: bem
          split: test
        metrics:
          - type: wer
            value: 37.96
            name: WER
      - task:
          type: automatic-speech-recognition
          name: Automatic Speech Recognition
        dataset:
          name: BembaSpeech
          type: BembaSpeech
          config: bem
          split: test
        metrics:
          - type: wer
            value: 37.96
            name: WER
      - task:
          type: automatic-speech-recognition
          name: Automatic Speech Recognition
        dataset:
          name: BembaSpeech
          type: BembaSpeech
          config: en
          split: test
        metrics:
          - type: wer
            value: 41.89
            name: WER

whisper-large-v3-bem

This model is a fine-tuned version of openai/whisper-large-v3 on the BembaSpeech bem dataset. It achieves the following results on the evaluation set:

  • Loss: 0.3448
  • Wer: 0.3758

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: 1.75e-05
  • train_batch_size: 4
  • eval_batch_size: 4
  • seed: 42
  • gradient_accumulation_steps: 6
  • total_train_batch_size: 24
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • lr_scheduler_warmup_steps: 500
  • num_epochs: 2.0
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Validation Loss Wer
0.4685 1.0084 500 0.3448 0.3758

Framework versions

  • Transformers 4.44.0
  • Pytorch 2.4.0+cu121
  • Datasets 2.21.0
  • Tokenizers 0.19.1