Kijai

/

LTX2.3_comfy

like 66

old one with new files (diffusion model, vae both video and audio), base on KJ screenshot.

I want a workflow too..

I want a workflow too..

Uploaded a simple workflow to test here https://huggingface.co/RuneXX/LTX-2.3-Workflows/
But you can really just use any and all workflows you want, such as ComfyUI default ones, and simply swap out the model loaders as per front page here : https://huggingface.co/Kijai/LTX2.3_comfy

THANKS! I have just downloaded this workflow and ran it on our RTX6000 Pro (96 gb VRAM). It can create a 10 second 1920x1088 video in 90 seconds! Quality seems to be quite good. (Wan 2.2 still has a better and higher of detail but takes MUCH longer to generate, more like 10-15 minutes for a similar video length and size)

Would you be able to modify the workflow and add start and end image support?

Would you be able to modify the workflow and add start and end image support?

Sure. Already added some for LTX-2.0.
will adapt them for LTX-2.3 ;-)

Thanks!

Would you be able to modify the workflow and add start and end image support?

Added for First-Middle-Last Frame and First-Last Frame (In-Place frame injection)
https://huggingface.co/RuneXX/LTX-2.3-Workflows/

(I'll also add some quite similar with image guide, instead of frame injection, that often can give a bit smoother results, letting the AI have a bit more freedom)

Cool - thank you!