whisper-finetune-th

This model is a fine-tuned version of biodatlab/whisper-th-medium-combined on the common_voice_17_0 dataset. It achieves the following results on the evaluation set:

• Loss: 0.1015
• Wer: 15.0453
• Cer: 3.8830

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: 16
• eval_batch_size: 8
• 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: 4000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer	Cer
0.2829	0.4873	1000	0.1345	20.0856	5.3644
0.1548	0.9747	2000	0.1161	17.6348	4.5783
0.1775	1.4620	3000	0.1074	15.9448	4.1193
0.1477	1.9493	4000	0.1015	15.0453	3.8830

Framework versions

• Transformers 4.44.2
• Pytorch 2.4.0+cu121
• Datasets 2.21.0
• Tokenizers 0.19.1