yosun
yosun
yeah comfy can get super wild - also video to 3d nodes too. iirc replicate has a 2 image limit as input. comfy AST supports N multimodal inputs - so...
exactly!
i don't understand the use case here... it's kind of useless to limit to 800 faces.
actually kind of surprised that kaggle or other sources don't have data sets of david's voice. youtube -> extract audio -> https://vocalremover.org/ -> spectrum analysis distill to 1 minute of...
actually there are several openesource models you can do voice cloning with - here's one from a few months ago trained on my acting voice https://replicate.com/p/jdmsp6jbaqrd5jqii7l3pav7s4
just a note that it's often a random partial string... an extra /g or an extra /grad ... we used to have to custom string replace each space's gradio api,...
@atteneder can you help?
...
hmm so most of the above works in editor but it seems assigning shorthand does not work on device or perhaps instantiation of new gameobject doesn't work? ```` Exception: chunk:6:...