metadata
license: apache-2.0
tags:
- generated_from_trainer
metrics:
- accuracy
model-index:
- name: wav2vec2-base-EMOPIA
results: []
wav2vec2-base-EMOPIA
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: 1.0286
- Accuracy: 0.6200
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: 1e-05
- train_batch_size: 4
- eval_batch_size: 4
- seed: 42
- gradient_accumulation_steps: 3
- total_train_batch_size: 12
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 15
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|---|---|---|---|---|
| 1.2694 | 0.98 | 41 | 1.1830 | 0.4800 |
| 1.0457 | 1.98 | 82 | 1.0959 | 0.4400 |
| 0.9201 | 2.98 | 123 | 0.9280 | 0.5600 |
| 0.8003 | 3.98 | 164 | 0.9346 | 0.7000 |
| 0.7616 | 4.98 | 205 | 0.9257 | 0.6600 |
| 0.6955 | 5.98 | 246 | 0.8411 | 0.7200 |
| 0.6743 | 6.98 | 287 | 0.9604 | 0.6200 |
| 0.5737 | 7.98 | 328 | 0.8471 | 0.7000 |
| 0.5269 | 8.98 | 369 | 1.0581 | 0.6000 |
| 0.4671 | 9.98 | 410 | 1.0365 | 0.6400 |
| 0.4227 | 10.98 | 451 | 1.0994 | 0.6000 |
| 0.3688 | 11.98 | 492 | 1.0964 | 0.5800 |
| 0.355 | 12.98 | 533 | 1.0586 | 0.5800 |
| 0.3232 | 13.98 | 574 | 1.0414 | 0.6000 |
| 0.3045 | 14.98 | 615 | 1.0286 | 0.6200 |
Framework versions
- Transformers 4.18.0
- Pytorch 1.10.1+cu102
- Datasets 2.4.0
- Tokenizers 0.12.1