https://github.com/tomfran/ravdess-ser
Speech emotion recognition on the RAVDESS dataset.
https://github.com/tomfran/ravdess-ser
data-augmentation machine-learning speech-emotion-recognition
Last synced: 11 months ago
JSON representation
Speech emotion recognition on the RAVDESS dataset.
- Host: GitHub
- URL: https://github.com/tomfran/ravdess-ser
- Owner: tomfran
- License: apache-2.0
- Created: 2022-06-08T14:33:38.000Z (about 4 years ago)
- Default Branch: main
- Last Pushed: 2022-08-23T09:21:52.000Z (almost 4 years ago)
- Last Synced: 2025-01-06T14:27:26.964Z (over 1 year ago)
- Topics: data-augmentation, machine-learning, speech-emotion-recognition
- Language: Jupyter Notebook
- Homepage:
- Size: 3.36 MB
- Stars: 1
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# Speech emotion recognition on RAVDESS
The goal of the project is to perform speech emotion recognition on audio files.
The dataset in use contains both songs and speech, by 24 distinct actors performing 8 different emotions.
## Code structure
All the code is organized inside `src` where you can find:
- `src/loader.py`: loading of raw data;
- `src/data_preprocessing.py`: feature extraction from raw data;
- `src/dataset.py`: dataset management, such as splits, scaling, etc;
- `src/models.py`: collection of functions that define or train models;
- `src/utility.py`: plots and initial study on the dataset.
## Notebooks
The study is divided among different notebooks:
- `01_exploration.ipynb`: initial study on the dataset;
- `02_training.ipynb`: initial training of neural networks and simpler models;
- `03_data_augmentation.ipynb`: experiments with data augmentation.
- `04_cross_validation.ipynb`: cross validation results.
[Report link](https://github.com/tomfran/ravdess-ser/blob/main/report/report.pdf)