thomas0104/whisper-large-v2-nan-tw-only-char

Whisper large-v2 nan-tw only char

This model is a fine-tuned version of openai/whisper-large-v2 on the mozilla-foundation/common_voice_11_0 nan-tw dataset. It achieves the following results on the evaluation set:

• Loss: 1.0351
• Wer: 45.3740
• Cer: 45.4573

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: 2
• eval_batch_size: 2
• 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
0.6011	1.04	1000	1.1100	55.0229	55.2068
0.1773	2.08	2000	1.2055	58.6565	58.7685
0.015	3.13	3000	1.0932	48.6412	48.8077
0.0131	5.01	4000	1.0531	45.7099	45.8497
0.0001	6.05	5000	1.0351	45.3740	45.4573

Framework versions

• Transformers 4.27.0.dev0
• Pytorch 1.13.1+cu117
• Datasets 2.8.0
• Tokenizers 0.13.2