xmzhu/whisper-small-zh-TW

Whisper Small Chinese (Taiwanese Mandarin)

This model is a fine-tuned version of xmzhu/whisper-small-zh on the mozilla-foundation/common_voice_11_0 zh-TW dataset. It achieves the following results on the evaluation set:

• Loss: 0.2639
• Wer: 42.9887

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: 500
• training_steps: 5000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.0058	6.02	1000	0.2445	43.0911
0.0006	13.02	2000	0.2639	42.9887
0.0003	20.01	3000	0.2787	43.1934
0.0002	27.0	4000	0.2877	43.5415
0.0002	33.02	5000	0.2910	43.5824

Framework versions

• Transformers 4.26.0.dev0
• Pytorch 1.13.1+cu117
• Datasets 2.7.1.dev0
• Tokenizers 0.13.2