[QUESTION] how to use batch size?

[targor]

Is this a new feature request?

• I have searched the existing issues

Wanted change

• How can i use the large-v2 model with fp16 and a batch size of 8 in docker compose? I do not see any documentation about flags for the bacth size, maybe it is not possible to set it?

• Is it possible to set and use vad models?

Reason for change

Maybe better performance regarding to the original documentation of faster-whisper

Proposed code change

No response

[github-actions]

Thanks for opening your first issue here! Be sure to follow the relevant issue templates, or risk having this issue marked as invalid.

[mg-dev25]

@targor Hi, if you want to play with this parameter you have to use directly the faster-whisper main project. This project only use the wyoming-faster-whisper to build it's image, and the ENV variable are only replacing this value:

parser.add_argument("--model", required=True, help="Name of faster-whisper model to use (or auto)") parser.add_argument("--beam-size", type=int, default=5, help="Size of beam during decoding (0 for auto)") parser.add_argument("--language", help="Default language to set for transcription")