[WIP] Add language selection for streaming with whisper + Improve tests #48
Conversation
Signed-off-by: AudranBert <[email protected]>
Signed-off-by: AudranBert <[email protected]>
Signed-off-by: AudranBert <[email protected]>
whisper/stt/processing/streaming.py
Outdated
| @@ -43,17 +44,20 @@ async def wssDecode(ws: WebSocketServerProtocol, model_and_alignementmodel): | |||
| try: | |||
| config = json.loads(res)["config"] | |||
| sample_rate = config["sample_rate"] | |||
| language = config.get("language", None) | |||
There was a problem hiding this comment.
Don't we need to update the doc, specifying the language can be specified in the configuration of the streaming request?
There was a problem hiding this comment.
Also, why not to support language="*" (as in the env variable for the language in offline decoding)
that would need to replace "*" by None when reading the config
There was a problem hiding this comment.
Don't we need to update the doc, specifying the language can be specified in the configuration of the streaming request?
Yes I need to update the doc
Signed-off-by: AudranBert <[email protected]>
Signed-off-by: AudranBert <[email protected]>
Signed-off-by: AudranBert <[email protected]>
Signed-off-by: AudranBert <[email protected]>
|
Could you clarify the list of supported languages? For example, does it include "en," "fr," etc.? On the LinTO side, we consistently use BCP-47 codes for language representation. |
Add language selection for streaming with whisper, by default it will take the language found in the env settings. But you can pass a language in the config when starting streaming.
The PR is also improving tests to add tests about languages. Also removing some useless ones in order to reduce testing duration.
It also adds the possibility to pass a language in the config in case of offline decoding. It will enable having a same model instance used for multiple languages instead of launching another Docker.