Plans for a multi actor version?
It's incredible how well this works!
But it seems to work for monologues only, correct? I mean, you can't use this for actual dialogue, since there's no way to generate two or more speakers, correct?
Do you plan to support multiple speakers?
no plans as of now ,
you can write a script which supports multiple reference and generate a podcast.
@mindplay I noticed when I set CFG 1 I already got a weird mix of actors! So maybe specifying in the prompt who talks each time could do the trick? I don't expect much but who knows
no plans as of now ,
you can write a script which supports multiple reference and generate a podcast.
well yes, but the actors won't "hear" each other, won't have the context of the conversation, situation, mood, etc.
was this trained on conversations, like radio plays, that sort of thing? did you need to split the recordings by speaker first? I'm wondering where you even source recordings like these - of just one person speaking with emotion? where in media does that happen?