adding support for multiple languages

harvard-edge / dataperf-speech-example

Example workflow for our data-centric speech benchmark

17 stars 11 forks source link

adding support for multiple languages #23

Closed colbybanbury closed 1 year ago

colbybanbury commented 1 year ago

Adds support for multiple languages (en, id, pt) by:

changing the download process to download and extract data for 3 languages
modify selection and eval to accept a language parameter
update README

Note: Currently MLCube only works for english. This is because MLCube only allows parameters to be files. The MLCube workflow has been deprioritized so this isn't on the critical path for users.

colbybanbury commented 1 year ago

Added the suggested changes