audio_kws_1019_ver2

This model is a fine-tuned version of Kkonjeong/wav2vec2-base-korean on the None dataset. It achieves the following results on the evaluation set:

• Loss: 3.0284
• Accuracy: 0.6154

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: 0.0001
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 32
• optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• num_epochs: 11
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	1.0	2	4.1202	0.0769
No log	2.0	4	3.8631	0.3077
No log	3.0	6	3.6660	0.3077
No log	4.0	8	3.5015	0.3077
No log	5.0	10	3.3840	0.3846
No log	6.0	12	3.2883	0.3846
No log	7.0	14	3.2052	0.6154
No log	8.0	16	3.1339	0.6154
No log	9.0	18	3.0802	0.6154
No log	10.0	20	3.0445	0.6154
No log	11.0	22	3.0284	0.6154

Framework versions

• Transformers 5.18.0
• Pytorch 2.11.0+cu130
• Datasets 4.8.5
• Tokenizers 0.23.2