I have reinstalled the computer system and installed the dependencies. I have two computers with the
I have reinstalled the computer system and installed the dependencies. I have two computers with the same problem
![Digital [Starburst]](https://cdn.discordapp.com/avatars/220961909447458817/8ecbbb3132fc6e44312f3d6d2ae3df69.webp?size=40)
DI'm not getting the impression that SD is really built for video work. I've generated some mp4's I sent my family that made them motion sick. Other than that I haven't gotten it to do much.
Fi think it is because
Fyour folder has space character
DThe idea is cool, but I don't think the technology is there yet.
Fi will dm you now
Fi sent you gradio file put into this folder
![Digital [Starburst]](https://cdn.discordapp.com/avatars/220961909447458817/8ecbbb3132fc6e44312f3d6d2ae3df69.webp?size=16)
Fye not very good
DI decided I didn't have the patience to wait nine hours to see what happens with a ten second video. When I checked the benchmark rank for graphics cards about the only thing better than what I have was a 4080. It had more VRAM, but cost nearly double the 4070ti.
DBefore video generation can really be used for anything useful they are going to have figure out a way to speed it up somehow with lighter hardware.
Fye it is slow
ADo the following work for both dreambooth, Lora, Textual inversion:
Basic Captioning
BLIP Captioning
GIT Captioning
WED14 Captioning
Manual Captioning
Or are the specific to each?
Basic Captioning
BLIP Captioning
GIT Captioning
WED14 Captioning
Manual Captioning
Or are the specific to each?
Fyes all works
Fit just different results
Aare there strength and weaknesses for them or do most just use BLIP?
Fyes each one fits
Fdifferent cases
Fi try them
JHello Dr. Furkan, i followed your tutorial to get a lora trained using a rtx 4090 gpu, but at my studio we are wanting to use multi-gpu training for sdxl, ive run into issues trying to configure kohya ss to do this properly (it ends up running into a memory issue after maxing out the memory of the first gpu), would you be willing to help me with this issue?
Fi am giving private consultation and i can help
Fbut gpus are on same machine right?
Jyea theyre on the same machine
Fthen i can help
JThank you! when can we talk about it?
Fi am available for several hours more from now on
JJ@Furkan Gözükara SECourses These are the configure settings I use, am I doing anything wrong here to where I will hit a memory road block while training?
Flooking good
J@Furkan Gözükara SECourses here are my training settings
Fsadly i dont have best settings for SD 1.5
Ji was using your settings for sdxl lora training
Fah yes they are decent but not very best
Fbest is SDXL DreamBooth
Joh i see
JFyour install selections
For better
Fdelete your accelerator settings
Fit shows at the end where it is saved
Fdelete this
Jill try that, thank you
J@Furkan Gözükara SECourses hmm i still seem to be getting the error
Fyou have 3 gpus?
Jyes
Fwell your setup wrong
Fi can professionally help
For you can limit to 1 gpu
Feasily
Fset CUDA_VISIBLE_DEVICES=0
Fadd this to .bat file