This repository contains scripts for processing various emotional speech datasets, and training machine learning models on these datasets.
github: https://github.com/Broad-AI-Lab/emotion
Datasets
Currently the following datasets are supported, along with their respective directories in the datasets
directory:
cafe
– CaFEcrema-d
– CREMA-Ddemos
– DEMoSemodb
– EMO-DBemofilm
– EmoFilmenterface
– eNTERFACEiemocap
– IEMOCAPjl
– JL-corpusmsp-improv
– MSP-IMPROVportuguese
– Portuguese corpusravdess
– RAVDESSsavee
– SAVEEsemaine
– SEMAINEshemo
– ShEMOsmartkom
– SmartKomtess
– TESS
Some of these datasets, (e.g. IEMOCAP), have more specific preprocessing scripts that must be run before the general preprocessing scripts, in order to generate the annotations, transcripts, etc.
Recent Comments