how do i load a lora model on comfyui
how do i load a lora model on comfyui
A@Furkan Gözükara SECourses It may be helpful to you if you create channels for the different GPU Cards people have, like 20series, 30series, 40 series, 16series, that way people can confirm if your updates worked on their graphics cards, and also people with those graphic cards can go see if they should update or wait...
A
MBring it in after the model but before clip. I can’t remember off the top of my head but I think the node is under loader and then Lora.
Zyeah i don't get it, i make a model and when i activate it, it's completely different
Zwhen are they going to fix this program
MSo it worked normally?
M
Mthat’s where you put the lora.
Fye could be. the problem is i dont have card to test :/
MI also think you need a trigger word for Lora depending on how well or even over trained it is.
MIf it’s way over trained probably no trigger word is necessary.
MI recommend taking a brief look at Kohya and Comfyui. Both are excellent for different things and compliment automatic1111 brilliantly. The fact I have different outputs with the same prompts, models, Lora and sampler is also interesting. Would love to see you give comfyui a go with your images. And maybe try training them using Kohya too.
MJust to see how you found it.
Fye i should cover
Fbut first i plan a comprehensive lora training of web ui
Moh no need for a video. Just would love to hear what you think.
Fi really need rtx 3090 :/
MHaha. Don’t we all
MI have 3080 though so can’t complain.
Fexcept hand :d
Falmost perfect
Fmaking some images to put into video talking parts
Fi hope this video get viewed so many times :d
MThat’s brilliant. Well done. Are you just playing with prompts now to get it right?
Fvery distant shot
Fi trained a style
F
Fthat ensures consistency
Fprompt is almost same simple
Zcrashes on the lora loader
Z
Z
Ftoo complicated for majority
MInteresting. I have never had an issue with comfy reading loras.
MYeah. You need some understanding of node based ui’s. Hugely powerful though. And in a way it’s art of its own before you even hit queue
F
Zcan you send me a screenshot of your workflow so i can get an idea of what i'm doing wrong Major?
MI can’t right now as I am not actually anywhere near my computer. But I will at some point.
MIt all looks good to me though.
MThe only difference I have is I use safetensor not pt.
MTry downloading a Lora from somewhere and throwing that in to see if it works.
MLet me check the screenshot I sent earlier that I stole from another discord. If that’s got a complete workflow I’ll post that.
Zyeah i need to figure out what i need to output latent from my ksampler to
M
MThis has controlnet so you can ignore that bit.
Zyeah i used a different lora that was safetensors and worked fine
Zi'll see if i can generate a checkpoit from my lora in auto1111 and go from there
Zi can't generate a checkpoint cause I get this error Exception compiling checkpoint: The size of tensor a (4) must match the size of tensor b (320) at non-singleton dimension 1
MThat beyond my level of expertise I am afraid.
Zso i generate a lora that doesn't look like my model, and then can't convert it to a checkpoint either