README.md

metadata

license: apache-2.0
base_model: openai/whisper-large-v3
tags:
  - generated_from_trainer
datasets:
  - DewiBrynJones/commonvoice_18_0_cy
metrics:
  - wer
model-index:
  - name: whisper-large-v3-ft-cv-cy
    results:
      - task:
          name: Automatic Speech Recognition
          type: automatic-speech-recognition
        dataset:
          name: DewiBrynJones/commonvoice_18_0_cy default
          type: DewiBrynJones/commonvoice_18_0_cy
          args: default
        metrics:
          - name: Wer
            type: wer
            value: 0.2026720744363593

whisper-large-v3-ft-cv-cy

This model is a fine-tuned version of openai/whisper-large-v3 on the DewiBrynJones/commonvoice_18_0_cy default dataset. It achieves the following results on the evaluation set:

• Loss: 0.4426
• Wer: 0.2027

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
0.0611	4.0161	1000	0.3231	0.2272
0.0088	8.0321	2000	0.3693	0.2145
0.0008	12.0482	3000	0.4037	0.2050
0.0005	16.0643	4000	0.4351	0.2050
0.0003	20.0803	5000	0.4426	0.2027

Framework versions

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