wav2vec2-base-kscg-gender-classification
This model is a fine-tuned version of facebook/wav2vec2-base on the kscg_small_20v50 dataset. It achieves the following results on the evaluation set:
- Loss: 0.3592
- Accuracy: 0.8703
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: 512
- eval_batch_size: 64
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 10
- mixed_precision_training: Native AMP
Training results
Training Loss | Epoch | Step | Validation Loss | Accuracy |
---|---|---|---|---|
0.4652 | 1.0 | 105 | 0.4543 | 0.8202 |
0.2969 | 2.0 | 210 | 0.4235 | 0.8127 |
0.248 | 3.0 | 315 | 0.3742 | 0.8508 |
0.2244 | 4.0 | 420 | 0.3805 | 0.8592 |
0.1983 | 5.0 | 525 | 0.3535 | 0.8535 |
0.1992 | 6.0 | 630 | 0.3330 | 0.8658 |
0.1806 | 7.0 | 735 | 0.3836 | 0.8633 |
0.1628 | 8.0 | 840 | 0.3493 | 0.8738 |
0.1574 | 9.0 | 945 | 0.3418 | 0.8765 |
0.1483 | 10.0 | 1050 | 0.3592 | 0.8703 |
Framework versions
- Transformers 4.28.1
- Pytorch 1.12.1
- Datasets 2.11.0
- Tokenizers 0.12.1
- Downloads last month
- 54