search

LostRuins / koboldcpp

A simple one-file way to run various GGML and GGUF models with KoboldAI's UI
https://github.com/lostruins/koboldcpp
GNU Affero General Public License v3.0
4.34k stars 310 forks source link

Feature request: Ability to "Generate More" when using Speech Control #903

Closed NGgmnBQGwR closed 1 week ago

NGgmnBQGwR commented 3 weeks ago

In "koboldcpp-1.67", a great feature was added - "Speech Control". But the button that acts as a voice indicator or Push-To-Talk toggle is the same "Generate More" button that is available without this feature enabled.

However, that means that right now there is no way to make the model generate more text with enabled speech control? My request is to separate the functionality into 2 buttons.

That is, instead of either "Generate More": image OR Speech Control: image To have both buttons available together: image

LostRuins commented 3 weeks ago

I've added a workaround for this for the next version - you can try it on the experimental branch or connecting via https://lite.koboldai.net/

If you tap lightly on the button, it will generate more. If you press and hold, then PTT will be activated. Try it and let me know what you think.

NGgmnBQGwR commented 3 weeks ago

Thank you! I've tried the "concedo_experimental" branch, and it works, but it's not really obvious - other users will probably miss this feature.

LostRuins commented 3 weeks ago

I think they'll try tapping the button when they want to generate same as usual non-voice mode, so it should work

LostRuins commented 1 week ago

The fix is now in the latest released version, please try it.