V2V workflow?
Do you know of a way to do a basic V2V which enhances the input video that doesn't require a first-frame condition frame?
I've been trying to work on one, but without the conditional frame, the results either aren't as good/lacking in details, or the output resembles the input to a degree but are altered.
Yes, the "foley" one already have an optional part like that (enhances a video input).
But will make that into a regular V2V Video Enhance workflow as well, and upload asap
Oh I see, I hadn't checked out that workflow. I did do a test and it works great, thanks! I hope by LTX v2.5 the motion blur issue is finally no more.
Btw any ideas why the Multimodal Guider node, which is supposed to be better than the regular CFG node, doesn't work with the Looping Sampler node?
I get this error
It something you added yourself I guess. In theory it should work.
Will try here.
Keep in mind they are only useful if you are NOT using distilled model and NOT using dev model with distilled lora (aka for both, CFG is 1).
The multi-modal nodes are for really slow workflows where you use Dev model at full steps (20-30 steps) and CFG is higher than 1
Thanks yes I was curious and replaced the CFG to see if it would make any difference. Btw I was wondering what are you using to set the colors of your nodes?
Btw I was wondering what are you using to set the colors of your nodes?
Good question. You mean just at the title bar? Or color in general.
In general you can right-click any node, and choose color at the menu...
For just coloring the title bar, its some node plugin that allows that. I will try see if i can remember which one if thats what you meant
For just coloring the title bar, its some node plugin that allows that. I will try see if i can remember which one if thats what you meant
Thanks, they look much nicer, since you seem to be able to use colors beyond the default 10 colors.
