<!-- 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. -->
wav2vec2-finetuned-on-shEMO
This model is a fine-tuned version of facebook/wav2vec2-base on the None dataset. It achieves the following results on the evaluation set:
- Loss: 1.0218
- Accuracy: 0.8
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: 8
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 32
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 20
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|---|---|---|---|---|
| 1.4581 | 1.0 | 75 | 1.4405 | 0.5767 |
| 1.0505 | 2.0 | 150 | 0.9797 | 0.71 |
| 0.9486 | 3.0 | 225 | 0.8445 | 0.74 |
| 0.7795 | 4.0 | 300 | 0.9015 | 0.6867 |
| 0.6058 | 5.0 | 375 | 0.7416 | 0.7767 |
| 0.5169 | 6.0 | 450 | 0.7565 | 0.78 |
| 0.4251 | 7.0 | 525 | 0.6422 | 0.82 |
| 0.3567 | 8.0 | 600 | 0.5284 | 0.8367 |
| 0.2806 | 9.0 | 675 | 0.6506 | 0.8033 |
| 0.2108 | 10.0 | 750 | 0.6477 | 0.8333 |
| 0.1468 | 11.0 | 825 | 0.5919 | 0.85 |
| 0.1624 | 12.0 | 900 | 0.6010 | 0.8533 |
| 0.1021 | 13.0 | 975 | 0.6798 | 0.8533 |
| 0.0647 | 14.0 | 1050 | 0.7265 | 0.8567 |
| 0.0502 | 15.0 | 1125 | 0.6910 | 0.8667 |
| 0.0326 | 16.0 | 1200 | 0.7374 | 0.8667 |
| 0.0554 | 17.0 | 1275 | 0.7250 | 0.8567 |
| 0.0312 | 18.0 | 1350 | 0.7943 | 0.8567 |
| 0.0729 | 19.0 | 1425 | 0.7315 | 0.86 |
| 0.0562 | 20.0 | 1500 | 0.7602 | 0.8533 |
Framework versions
- Transformers 4.28.1
- Pytorch 2.0.0+cu118
- Datasets 2.11.0
- Tokenizers 0.13.3
