wav2vec2-xlsr-korean-speech-emotion-recognition2_data_rebalance
This model is a fine-tuned version of jungjongho/wav2vec2-large-xlsr-korean-demo-colab_epoch15 on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.0124
- Accuracy: 0.9976
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: 2
- eval_batch_size: 2
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 4
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 2
- mixed_precision_training: Native AMP
Training results
Training Loss | Epoch | Step | Validation Loss | Accuracy |
---|---|---|---|---|
1.1219 | 0.04 | 200 | 1.4543 | 0.5792 |
0.7585 | 0.08 | 400 | 0.6959 | 0.7301 |
0.5959 | 0.13 | 600 | 0.7356 | 0.7671 |
0.4048 | 0.17 | 800 | 0.2661 | 0.9128 |
0.3101 | 0.21 | 1000 | 0.2470 | 0.9291 |
0.2805 | 0.25 | 1200 | 0.2115 | 0.9394 |
0.2885 | 0.29 | 1400 | 0.3597 | 0.9152 |
0.1703 | 0.33 | 1600 | 0.3973 | 0.9142 |
0.1818 | 0.38 | 1800 | 0.2164 | 0.9543 |
0.1727 | 0.42 | 2000 | 0.1133 | 0.9730 |
0.1038 | 0.46 | 2200 | 0.0826 | 0.9851 |
0.1343 | 0.5 | 2400 | 0.0823 | 0.9820 |
0.1412 | 0.54 | 2600 | 0.0762 | 0.9830 |
0.1321 | 0.58 | 2800 | 0.0786 | 0.9806 |
0.0738 | 0.63 | 3000 | 0.1036 | 0.9810 |
0.0998 | 0.67 | 3200 | 0.1984 | 0.9640 |
0.1135 | 0.71 | 3400 | 0.0775 | 0.9841 |
0.0552 | 0.75 | 3600 | 0.0923 | 0.9827 |
0.0633 | 0.79 | 3800 | 0.0518 | 0.9900 |
0.0769 | 0.83 | 4000 | 0.0599 | 0.9875 |
0.1026 | 0.88 | 4200 | 0.0800 | 0.9841 |
0.0641 | 0.92 | 4400 | 0.2396 | 0.9606 |
0.1068 | 0.96 | 4600 | 0.0653 | 0.9875 |
0.0802 | 1.0 | 4800 | 0.0844 | 0.9855 |
0.0483 | 1.04 | 5000 | 0.0984 | 0.9834 |
0.0392 | 1.09 | 5200 | 0.1092 | 0.9813 |
0.0408 | 1.13 | 5400 | 0.0719 | 0.9900 |
0.0388 | 1.17 | 5600 | 0.0494 | 0.9903 |
0.0253 | 1.21 | 5800 | 0.1486 | 0.9751 |
0.0448 | 1.25 | 6000 | 0.1370 | 0.9782 |
0.0415 | 1.29 | 6200 | 0.0508 | 0.9907 |
0.0552 | 1.34 | 6400 | 0.0332 | 0.9941 |
0.065 | 1.38 | 6600 | 0.0479 | 0.9900 |
0.0391 | 1.42 | 6800 | 0.0470 | 0.9910 |
0.0339 | 1.46 | 7000 | 0.0550 | 0.9886 |
0.0525 | 1.5 | 7200 | 0.0389 | 0.9920 |
0.0393 | 1.54 | 7400 | 0.0543 | 0.9910 |
0.0488 | 1.59 | 7600 | 0.0205 | 0.9965 |
0.0253 | 1.63 | 7800 | 0.0240 | 0.9948 |
0.0438 | 1.67 | 8000 | 0.0308 | 0.9952 |
0.0291 | 1.71 | 8200 | 0.0160 | 0.9969 |
0.0235 | 1.75 | 8400 | 0.0124 | 0.9969 |
0.0061 | 1.8 | 8600 | 0.0191 | 0.9962 |
0.022 | 1.84 | 8800 | 0.0178 | 0.9958 |
0.0176 | 1.88 | 9000 | 0.0135 | 0.9965 |
0.0168 | 1.92 | 9200 | 0.0161 | 0.9969 |
0.0068 | 1.96 | 9400 | 0.0124 | 0.9976 |
Framework versions
- Transformers 4.22.0.dev0
- Pytorch 1.12.1+cu113
- Datasets 2.4.1.dev0
- Tokenizers 0.12.1
- Downloads last month
- 90