Indonesian Speech Emotion Recognition By Fine-Tuning Wav2Vec 2.0

The model is a fine-tuned version of facebook/wav2vec2-base for a Speech Emotion Recognition (SER) task.

The dataset used to fine-tune the original pre-trained model is the RAVDESS dataset. This dataset provides 1440 samples of recordings from actors performing on 5 different emotions in Bahasa Indonesia, which are:

emotions = ['angry', 'disgust', 'fear', 'happy', 'sad']

It achieves the following results on the evaluation set:

• Loss: 0.5023
• Accuracy: 0.8223

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-5
• train_batch_size: 32
• eval_batch_size: 32
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 10
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 10
• mixed_precision_training: Native AMP

Training results

Citation

@misc {alianur_rahman_2024,
	author       = { {Alianur Rahman} },
	title        = { wav2vec2-base-indonesian-speech-emotion-recognition (Revision 1fcfcf1) },
	year         = 2024,
	url          = { https://huggingface.co/alianurrahman/wav2vec2-base-indonesian-speech-emotion-recognition }
}

Contact

Any doubt, contact me on Twitter.

Framework versions

• Transformers 4.45.1
• Pytorch 2.2.2
• Datasets 3.0.1
• Tokenizers 0.20.0