The docs on audioTimestamp say:
“Optional. If enabled, audio timestamp will be included in the request to the model.”
What does exactly this mean? Do we need to provide timestamps as a part of the prompt, alongside with the audio?
Or is it an instruction to the LLM to generate timestamps?
Thanks!
Hi @noseratio,
Welcome to the Forum,
Enabling audioTimestamp means that the model will include audio timestamp information in its response. It does not mean you need to provide timestamps as part of your prompt alongside the audio, nor is it an instruction for the LLM to generate timestamps from scratch if you haven’t provided audio. Rather, it suggests that if you provide audio, the model will process it and output timestamps as part of its generated response.
Thank you!