This repository contains scripts for processing emotional speech datasets, and training machine learning models on the datasets. The code was developed mainly to facilitate my PhD research, but I have tried to make the Python library more generally useful and full of utilities for processing datasets and training/testing models.
See datsets/README.md for more information about
the supported datasets and the required processing.
This is a Python library with utilities for processing (emotional) speech datasets and training/testing models. There are also associated scripts for individual tasks.
This project requires Python 3.7+. It is advised to run the scripts in a Python virtual environment. One can be created with the command
python -m venv .venv
Then you can use this virtual environment:
. .venv/bin/activate
This project has a number of Python dependencies, including NumPy, TensorFlow, PyTorch, scikit-learn and Pandas. You should run
pip install -r requirements.txt
to install all the Python dependencies.
To install the ertk library:
python setup.py install
Or, if you want to develop continuously:
python setup.py develop
Note that this will not install the scripts.
Some third-party tools are included in this repository, under the
third_party directory. See the README for
more details.
Scripts are contained in the scripts/ directory, in a
subdirectory roughly corresponding to their function. Each script should
have usage info if you give --help as an argument.
The papers/ directory contains copies of scripts used to
run experiments and results for a given paper that we've published. Each
directory corresponds to a publicaiton and has a README with a
description of the paper and how to run the experiments. Each
publication will also be associated with a git tag and a GitHub release
on the releases
page.
NOTE: Scripts in the papers/ directory will probably not be updated.
Please checkout the appropriate git tag to use these scripts.