jethrowang/vanilla-whisper-tiny

Whisper Tiny Hakka Condenser

This model is a fine-tuned version of openai/whisper-tiny on the HAT ASR Aligned dataset. It achieves the following results on the evaluation set:

• Loss: 0.1729
• Cer: 10.2307

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: 64
• eval_batch_size: 32
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 1521
• training_steps: 15210
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Cer
0.2476	0.9993	1521	0.4437	23.6551
0.0892	1.9987	3042	0.2482	14.6693
0.0543	2.9980	4563	0.2007	11.1774
0.0361	3.9974	6084	0.1847	12.4939
0.0235	4.9967	7605	0.1791	10.5405
0.0157	5.9961	9126	0.1727	10.9000
0.0121	6.9954	10647	0.1724	11.1554
0.0082	7.9947	12168	0.1720	10.3694
0.0059	8.9941	13689	0.1732	10.4053
0.0049	9.9934	15210	0.1729	10.2307

Framework versions

• Transformers 4.42.3
• Pytorch 2.3.0+cu121
• Datasets 2.20.0
• Tokenizers 0.19.1