<!-- 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. -->
asd_pron_w2v_reg_balanced_500_79_corr_converse
This model is a fine-tuned version of slplab/wav2vec2-xls-r-300m_phone-mfa_korean on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.5163
- Spearman Correlation: 0.1648
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: 3e-05
- train_batch_size: 16
- eval_batch_size: 16
- seed: 42
- gradient_accumulation_steps: 8
- total_train_batch_size: 128
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 30
- mixed_precision_training: Native AMP
Training results
Training Loss | Epoch | Step | Validation Loss | Spearman Correlation |
---|---|---|---|---|
0.5482 | 1.0 | 308 | 0.5296 | 0.0806 |
0.3928 | 2.0 | 616 | 0.5051 | 0.1624 |
0.3876 | 3.0 | 924 | 0.5314 | 0.2300 |
0.3913 | 4.0 | 1232 | 0.5007 | 0.2124 |
0.3896 | 5.0 | 1540 | 0.5150 | 0.1451 |
0.3894 | 6.0 | 1848 | 0.5103 | 0.1414 |
0.3896 | 7.0 | 2156 | 0.5173 | 0.2652 |
0.3909 | 8.0 | 2464 | 0.5151 | 0.2192 |
0.3875 | 9.0 | 2772 | 0.5093 | 0.1937 |
0.3873 | 10.0 | 3080 | 0.5134 | 0.1517 |
0.3938 | 11.0 | 3388 | 0.5084 | 0.3055 |
0.3894 | 12.0 | 3696 | 0.5025 | 0.1990 |
0.3907 | 13.0 | 4004 | 0.5249 | 0.2072 |
0.389 | 14.0 | 4312 | 0.5157 | 0.2459 |
0.3887 | 15.0 | 4620 | 0.5033 | 0.2882 |
0.3882 | 16.0 | 4928 | 0.5078 | 0.2449 |
0.3906 | 17.0 | 5236 | 0.5215 | 0.2655 |
0.3903 | 18.0 | 5544 | 0.5101 | 0.2863 |
0.3905 | 19.0 | 5852 | 0.5090 | 0.2158 |
0.3877 | 20.0 | 6160 | 0.5226 | 0.1800 |
0.3874 | 21.0 | 6468 | 0.5123 | 0.2398 |
0.3901 | 22.0 | 6776 | 0.5108 | 0.2126 |
0.3913 | 23.0 | 7084 | 0.5195 | 0.1674 |
0.3857 | 24.0 | 7392 | 0.5099 | 0.2044 |
0.387 | 25.0 | 7700 | 0.5174 | 0.2156 |
0.3878 | 26.0 | 8008 | 0.5202 | 0.1925 |
0.3877 | 27.0 | 8316 | 0.5119 | 0.1712 |
0.389 | 28.0 | 8624 | 0.5172 | 0.1732 |
0.3909 | 29.0 | 8932 | 0.5195 | 0.1783 |
0.3876 | 30.0 | 9240 | 0.5163 | 0.1648 |
Framework versions
- Transformers 4.13.0
- Pytorch 2.0.0+cu118
- Datasets 2.14.4
- Tokenizers 0.10.3