| --- |
| library_name: transformers |
| base_model: Kkonjeong/wav2vec2-base-korean |
| tags: |
| - generated_from_trainer |
| metrics: |
| - accuracy |
| model-index: |
| - name: korean_kws2 |
| results: [] |
| --- |
| |
| <!-- This model card has been generated automatically according to the information the Trainer had access to. You |
| should probably proofread and complete it, then remove this comment. --> |
|
|
| # korean_kws2 |
| |
| This model is a fine-tuned version of [Kkonjeong/wav2vec2-base-korean](https://huggingface.co/Kkonjeong/wav2vec2-base-korean) on the None dataset. |
| It achieves the following results on the evaluation set: |
| - Loss: 0.3954 |
| - Accuracy: 0.9474 |
| |
| ## 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: 30 |
| - mixed_precision_training: Native AMP |
| |
| ### Training results |
| |
| | Training Loss | Epoch | Step | Validation Loss | Accuracy | |
| |:-------------:|:-----:|:----:|:---------------:|:--------:| |
| | No log | 1.0 | 2 | 1.9452 | 0.0526 | |
| | No log | 2.0 | 4 | 1.9039 | 0.0 | |
| | No log | 3.0 | 6 | 1.7879 | 0.1053 | |
| | No log | 4.0 | 8 | 1.7879 | 0.1053 | |
| | No log | 5.0 | 10 | 1.6095 | 0.4211 | |
| | No log | 6.0 | 12 | 1.4181 | 0.8947 | |
| | No log | 7.0 | 14 | 1.3656 | 0.8947 | |
| | No log | 8.0 | 16 | 1.2083 | 0.8947 | |
| | No log | 9.0 | 18 | 1.1409 | 0.9474 | |
| | No log | 10.0 | 20 | 1.0315 | 0.9474 | |
| | No log | 11.0 | 22 | 0.9487 | 0.9474 | |
| | No log | 12.0 | 24 | 0.8729 | 0.9474 | |
| | No log | 13.0 | 26 | 0.7926 | 0.9474 | |
| | No log | 14.0 | 28 | 0.7395 | 0.9474 | |
| | No log | 15.0 | 30 | 0.6722 | 0.9474 | |
| | No log | 16.0 | 32 | 0.6302 | 0.9474 | |
| | No log | 17.0 | 34 | 0.5891 | 0.9474 | |
| | No log | 18.0 | 36 | 0.5686 | 0.9474 | |
| | No log | 19.0 | 38 | 0.5241 | 0.9474 | |
| | No log | 20.0 | 40 | 0.5047 | 0.9474 | |
| | No log | 21.0 | 42 | 0.4811 | 0.9474 | |
| | No log | 22.0 | 44 | 0.4597 | 0.9474 | |
| | No log | 23.0 | 46 | 0.4416 | 0.9474 | |
| | No log | 24.0 | 48 | 0.4274 | 0.9474 | |
| | No log | 25.0 | 50 | 0.4168 | 0.9474 | |
| | No log | 26.0 | 52 | 0.4093 | 0.9474 | |
| | No log | 27.0 | 54 | 0.4038 | 0.9474 | |
| | No log | 28.0 | 56 | 0.3999 | 0.9474 | |
| | No log | 29.0 | 58 | 0.3968 | 0.9474 | |
| | No log | 30.0 | 60 | 0.3954 | 0.9474 | |
| |
| |
| ### Framework versions |
| |
| - Transformers 5.0.0 |
| - Pytorch 2.10.0+cu128 |
| - Datasets 4.0.0 |
| - Tokenizers 0.22.2 |
| |
