Will Speaker Label be Required for OpenAudio text2sementic module Training?

[LEECHOONGHO]

When I looked into the inference code, The Input for text2sementic module is like

{||Speaker0||} {Prompt Texts} {Prompt Audio Tokens} {||IM_END||} {||Speaker0||} {Texts}

and generates {Audio Tokens} {||IM_END||}.

If the inference proceeds like this, would two voices from the same speaker be needed to train text2sementic model?
Speaker ID is used to extract two voices from the same speaker.

[PoTaTo-Mika]

speaker is not needed in current version.

[LEECHOONGHO]

Thank you for your reply!