Add more datasets into data (all publics) - Githubissues

felixbur / nkululeko

Machine learning speaker characteristics

MIT License

26 stars 4 forks source link

Add more datasets into data (all publics) #122

Closed bagustris closed 1 month ago

bagustris commented 1 month ago

This PR add more datasets:

jnv: Japanese non-verbal speech
jnvv: Japanese verbal and non-verbal speech
kbes: bengali speeh emotion
tth: Vietnamese speech emotion
banglaser: bengali speech emotion
polish: polish emotion speech
nemo: another polish emotion speech
emouerj: portuguese emotional speech
eaed: Egyptian Arabic emotional speech
kia: Korean emotional wake-up words
turev-db: turkish emotional speech
erysac: young russian emotional speech

Some information about these datasets can be found in this google docs.

This PR also adds available labels regardless target given in config or not (useful to check when error typing labels manually).