wav2vec2-lg-xlsr-en-speech-emotion-recognition-finetuned-ravdess-v8
This model is a fine-tuned version of ehcalabres/wav2vec2-lg-xlsr-en-speech-emotion-recognition on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.6778
- Accuracy: 0.75
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: 0.0001
- train_batch_size: 4
- eval_batch_size: 4
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 8
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 3
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|---|---|---|---|---|
| 2.0178 | 0.15 | 25 | 1.8431 | 0.6181 |
| 1.7082 | 0.31 | 50 | 1.5052 | 0.5833 |
| 1.4444 | 0.46 | 75 | 1.3458 | 0.5972 |
| 1.3888 | 0.62 | 100 | 1.2760 | 0.5972 |
| 1.1819 | 0.77 | 125 | 1.1075 | 0.6667 |
| 1.1615 | 0.93 | 150 | 1.0666 | 0.625 |
| 1.1659 | 1.08 | 175 | 1.3450 | 0.5694 |
| 0.9798 | 1.23 | 200 | 0.9866 | 0.6528 |
| 0.9893 | 1.39 | 225 | 0.9311 | 0.6806 |
| 0.9357 | 1.54 | 250 | 0.9783 | 0.6736 |
| 0.7998 | 1.7 | 275 | 0.7924 | 0.7014 |
| 0.7444 | 1.85 | 300 | 0.8980 | 0.6806 |
| 0.7648 | 2.01 | 325 | 0.8994 | 0.7153 |
| 0.607 | 2.16 | 350 | 0.9416 | 0.6597 |
| 0.5551 | 2.31 | 375 | 0.7791 | 0.7431 |
| 0.5495 | 2.47 | 400 | 0.7665 | 0.7431 |
| 0.5498 | 2.62 | 425 | 0.8017 | 0.7222 |
| 0.4887 | 2.78 | 450 | 0.6967 | 0.7639 |
| 0.5308 | 2.93 | 475 | 0.6857 | 0.7569 |
Framework versions
- Transformers 4.32.1
- Pytorch 2.0.1+cu118
- Datasets 2.14.4
- Tokenizers 0.13.3