Thanks for the question, The models you’re seeing, DisplayName 20230915
and 20231026 Whisper Preview
, are indeed configured for the en-US
locale. This means they have been trained and optimized for English as spoken in the United States.
However, this doesn’t necessarily mean they won’t work for other locales. They can still process and understand English inputs from other locales, but the efficiency might vary as the language usage, accents, and other linguistic nuances can differ across locales.
For optimal performance, it’s generally recommended to use a model that matches your specific locale. If a model for your locale isn’t available, an en-US
model could be a reasonable alternative for English language inputs.