Specifying the language of the transcription #7

Open
opened 2024-06-20 10:29:38 +02:00 by erth9960 · 1 comment
Owner

Whisper sometimes gets confused about the language of a file. This means that we need to be able to optionally specify a language when submitting a job, and pass that language on to the backend.

The backend accepts an optional language parameter as a top-level key in the job specification.

Whisper sometimes gets confused about the language of a file. This means that we need to be able to optionally specify a language when submitting a job, and pass that language on to the backend. The backend accepts an optional `language` parameter as a top-level key in the job specification.
Author
Owner

The format of the language key is the two-letter ISO-639-1 language code. Real-life use will probably only see en and sv occur, but whisper can handle a number of languages, see here.

What subset to allow is unclear. One option would be to allow any language to be passed and let users deal with the consequences of their choice.

Whisper will auto-detect the language if the key is missing or has a null value. This should be the default.

The format of the language key is the two-letter [ISO-639-1](https://en.wikipedia.org/wiki/ISO_639-1) language code. Real-life use will probably only see `en` and `sv` occur, but whisper can handle a number of languages, see [here](https://github.com/openai/whisper). What subset to allow is unclear. One option would be to allow any language to be passed and let users deal with the consequences of their choice. Whisper will auto-detect the language if the key is missing or has a null value. This should be the default.
Sign in to join this conversation.
No Label
No Milestone
No project
No Assignees
1 Participants
Notifications
Due Date
The due date is invalid or out of range. Please use the format 'yyyy-mm-dd'.

No due date set.

Dependencies

No dependencies set.

Reference: DMC/whisper-frontend#7
No description provided.