xmzhu/whisper-tiny-zh-TW-baseline

Whisper Tiny Chinese (Taiwanese Mandarin)

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

• Loss: 0.4811
• Wer: 69.3552

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
0.1257	6.02	1000	0.4811	69.3552
0.0122	13.02	2000	0.5494	70.6858
0.0044	20.01	3000	0.5851	71.3818
0.0028	27.0	4000	0.6098	72.6919
0.0022	33.02	5000	0.6159	72.7943

Framework versions

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