ljs0710/whisper-small-finetuning-ko

whisper-small-for-filler_words

This model is a fine-tuned version of openai/whisper-small on the Junhoee/STT_Korean_Dataset_80000 dataset. It achieves the following results on the evaluation set:

• Loss: 0.4023
• Cer: 12.9583

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: 0.0002
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 16
• 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: 500
• training_steps: 4000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Cer
0.9422	0.25	1000	0.9523	42.9308
0.669	0.5	2000	0.6758	20.5051
0.5142	0.75	3000	0.4989	17.0488
0.3686	1.0	4000	0.4023	12.9583

Framework versions

• Transformers 4.52.0.dev0
• Pytorch 2.6.0+cu124
• Datasets 3.5.0
• Tokenizers 0.21.1