README.md

metadata

license: apache-2.0
base_model: openai/whisper-large-v3
tags:
  - generated_from_trainer
datasets:
  - DewiBrynJones/banc-trawsgrifiadau-bangor-clean-with-ccv
metrics:
  - wer
model-index:
  - name: whisper-large-v3-ft-btb-ccv-cy
    results:
      - task:
          name: Automatic Speech Recognition
          type: automatic-speech-recognition
        dataset:
          name: DewiBrynJones/banc-trawsgrifiadau-bangor-clean-with-ccv default
          type: DewiBrynJones/banc-trawsgrifiadau-bangor-clean-with-ccv
          args: default
        metrics:
          - name: Wer
            type: wer
            value: 0.29472481531011024

whisper-large-v3-ft-btb-ccv-cy

This model is a fine-tuned version of openai/whisper-large-v3 on the DewiBrynJones/banc-trawsgrifiadau-bangor-clean-with-ccv default dataset. It achieves the following results on the evaluation set:

• Loss: 0.4479
• Wer: 0.2947

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: 16
• eval_batch_size: 16
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 32
• 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
0.4118	0.7722	1000	0.4898	0.3627
0.2629	1.5444	2000	0.4272	0.3145
0.1677	2.3166	3000	0.4243	0.3039
0.0841	3.0888	4000	0.4493	0.2968
0.0845	3.8610	5000	0.4479	0.2947

Framework versions

• Transformers 4.44.0
• Pytorch 2.4.0+cu121
• Datasets 2.20.0
• Tokenizers 0.19.1