@Dr. Furkan Gözükara When decreasing ddim and alpha should I increase learning rate 5e-5 for 128 wha
@Dr. Furkan Gözükara When decreasing ddim and alpha should I increase learning rate 5e-5 for 128 what for example 64 ?
F@Dr. Furkan Gözükara I have 2 observations, I have the impression that the generation is long if I compare it to the one on openart in turbo mode. I took 2 A6000, btw, You said in your video that the comparison between lora and finetuning is like night and day. I wanted to know at what level, because if it's rendering quality, I think it's pretty similar. the first image is made with openart and the second with finetuning , it's even worse in finetuning because I've lost the comic illustration aspect.
F
Ffor info I was able to recover the comic illustration style by adding negative prompts, photorealistic, hyper-realistic, realistic details, etc.
FI think there's only one GPU working, whether it's for one image or 10 in a generation. is that normal? @Dr. Furkan Gözükara
D@Dr. Furkan Gözükara Any reliable way to convert
.pt .pth .bin into .safetensors?DGitHub
A simple repo for single or batch converting .pt, .pth & .bin models into .safetensors models. - MackinationsAi/convert-2-safetensors
D
GitHub
Converts pth/pt files to safetensors. Contribute to jtabox/safetensors-converter development by creating an account on GitHub.
DAnybody familiar with this error? Happens with both
.pth and .safetensors files.
D@Dr. Furkan Gözükara Any idea on how to resolve this error?
DGitHub
Collect super-resolution related papers, data, repositories - ChaofWang/Awesome-Super-Resolution
Fit is never added
Fplease leave a reply here
FGitHub
Allow yolo segmentation to sort by left-right, top-bottom, largest-smallest, and reverse that order.
Params added to Regional Segmentation Group
Params added to Regional Segmentation Group
Fdont change alpha and learning rate
Fwhat do you mean by openart?
Fyou need to manually add backend
Fshown in many tutorials by now
Fnot that i know
KOh it's working. You must have done something wrong installing
Fyou should ask the source - i dont know atm
F@Dr. Furkan Gözükara openart is a platform on which I used to create my images before they changed their prices. I used to make fairly high-quality lora flux, after a few image generations on Swarmui I find the quality very good, little waste, but it's a bit slow, I'm going to test with an L40 to go faster what do you think? would it be better with 2A6000?
Fyou can train 2x gpu
Fso yes almost 2x spee
Ffor fine tuning you cant you need 80 gb gpus
KMaybe you could get away with a quantized model of Flux, but I really coudn't say. I've tried Q5 and that gives rather good results. I'm not sure if it can be run with 8-steps lora though, so maybe there is no time savings either with that.
KThat PSU though 
F@Dr. Furkan Gözükara ok but for image creation and not model training? It's this slowness that bothers me
F2x gpu = 2x faster
Fsince you can generate 2x image at the same time
Fget 4x gpu = 4x speed 
KKnow the feeling. I'd rather create one image really fast than two images at double that time 
Fwell i generate more images to get best ones
KI should get a medal for my grammar 
Fwith flux usually everyone is good
KGenerating goes fast. It's the training that takes AGES!
Fmy question was to compare 2A6000 to 1L40 but I'll test it directly to see for myself, thanks again.
F2x a6000 would be faster
Fl40 has advantage in fp8 but that reduces quality when you apply architecture optimization
Fsee rtx 5090
Fgonna buy hopefully first day
FKMore like 
KAnyone know why ComfyUI does this?
KIt's non-destructive, but very irritating
K
Share and Run ComfyUI workflows in the cloud
DI don't think I did. It might just be that
JoyCaption needs way more performance than this?Fworking on CogVideoX1.5-5B-I2V quant
