Curiousfox/whisper_new_ver1

Hokkien-to-Tai Lo Whisper ver 1

This model is a fine-tuned version of openai/whisper-small on the Common Voice 12.0 dataset. It achieves the following results on the evaluation set:

• Loss: 0.8241
• Wer: 170.1861

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-06
• train_batch_size: 8
• eval_batch_size: 16
• 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: 1000
• training_steps: 8000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
1.4704	1.4620	1000	1.3168	199.1033
0.8377	2.9240	2000	0.9763	217.8659
0.6319	4.3860	3000	0.8829	215.1311
0.5061	5.8480	4000	0.8412	181.5736
0.4306	7.3099	5000	0.8232	173.9296
0.3712	8.7719	6000	0.8209	166.2408
0.3321	10.2339	7000	0.8229	165.9045
0.3198	11.6959	8000	0.8241	170.1861

Framework versions

• Transformers 4.48.3
• Pytorch 2.5.1+cu124
• Datasets 3.3.0
• Tokenizers 0.21.0