marcogdepinto/emotion-classification-from-audio-files
Deep learning classifier that predicts speaker emotions from audio files using RAVDESS and TESS datasets.
Not currently ranked — collecting fresh signals.
star history
This project trains neural networks on audio samples to classify 8 emotional states (neutral, calm, happy, sad, angry, fearful, disgust, surprised) from speech and song recordings. It uses the RAVDESS and TESS datasets totaling 5252 samples, processing audio with librosa for feature extraction and building models with Keras and TensorFlow. The classifier achieves an 80% F1 score across all emotion classes.
Frequently asked
- What is marcogdepinto/emotion-classification-from-audio-files?
- Deep learning classifier that predicts speaker emotions from audio files using RAVDESS and TESS datasets.
- Is emotion-classification-from-audio-files open source?
- Yes — marcogdepinto/emotion-classification-from-audio-files is open source, released under the GPL-3.0 license.
- What language is emotion-classification-from-audio-files written in?
- marcogdepinto/emotion-classification-from-audio-files is primarily written in Python.
- How popular is emotion-classification-from-audio-files?
- marcogdepinto/emotion-classification-from-audio-files has 433 stars on GitHub.
- Where can I find emotion-classification-from-audio-files?
- marcogdepinto/emotion-classification-from-audio-files is on GitHub at https://github.com/marcogdepinto/emotion-classification-from-audio-files.