https://www.patreon.com/posts/112099700
Patreon
Get more from SECourses: FLUX, Tutorials, Guides, Resources, Training, Scripts on Patreon
OJust a quick question on Supir upscaling: are there a collection of presets that match up with some of the tests run? I'm not seeing exactly how to configure this to match up with the grid tests run on the SECourses channel. Is there a go-to configuration that I can start with and experiment from there?
I'm likely most interested in upscaling images generated in Flux and likely 2x to 4x in resolution
I'm likely most interested in upscaling images generated in Flux and likely 2x to 4x in resolution
Fi have a default config
Fthat upscales with loyal to original image
Fand also we have Replicate.json
Fwhich is same as on replicate
OAh - is that in the lower right in this image?
Oare default and replicate different?
Fyep
Fpresets
Freplicate is same config on replicate
For almost same
OWhat does replicate do?
Oand this is using juggernaut-xl?
Ofor default?
Ook, i see. I need to use default and then actually load it.
OI incorrectly assumed default was the application default and it was already loaded.
Oty!@
Fcompare both
Fwith different base models
Fsee which you like most
Oty - i was just having some trouble getting started. much appreciated!
Elul so I was quite sure my RAM capacity being inadequate was the reason for my system basically being unusable for over a minute while Flux was loading stuff each time I changed LORAs etc... I had 32GB before... got the new kit in tonight, and this is what using Flux looks like so far:
E
EGlad I got 96GB & not 64GB, I would've had the same issue & that woulda pissed me off 
Cyou need vram, not ram
Cyou want everythign loaded into vram and on your gpu. not into ram or on your cpu
EI have an RTX 4090, the loading process before it's transferred to the GPU was where I was having hard lags, because I only had 32GB of system RAM and it was using ALL of it lol
Ci have a 4090 as well and don't have that sort of lag. what CPU do you have?
EHow much system RAM do you have? And my CPU is my next upgrade, right now I'm running a 7800X3D, the system was originally upgraded/built for gaming.
Cand you have an AMD - that's a problem. i have an intel. there are things you have to do for an AMD that you don't have to do for an intel. have you gone through CS1o's AMD guides?
ENope, and it's not having an issues now, it capping out system RAM with 32GB was the main issue I was having and that causing the system to basically be unusable for 2 minutes each time I changed the prompts or settings significantly. Are you running the full FP16 Flux Dev version?
Cyup. all the original models from BFL. i don't use the quantized stuff
Cor gguf
Ccan i DM you
Cit shouldn't be capping out system ram at all. there are some things you need to do to stop AMD from trying to do the GPU's job
Esure go ahead, I'm new to Flux so I have no idea what I'm doing 
and I don't know, I just know that every time I was loading new stuff it was hitting 99.9% of sytem RAM usage with 32GB and basically locking up the system for 2 minutes, and now it's only using 33-64GB depending on what I'm using with no system lag at all.
and I don't know, I just know that every time I was loading new stuff it was hitting 99.9% of sytem RAM usage with 32GB and basically locking up the system for 2 minutes, and now it's only using 33-64GB depending on what I'm using with no system lag at all.
D@Dr. Furkan Gözükara Hello dr, I have downloaded today two claimed de-distilled flux models, I didn't hade time to testing them, I don't know if kohya_ss is compatible con one of them, one of them is based on flux1-dev and the other in schnell, in one hand the distillation process as I have read is easier to remove on flux1-dev and it have better quality, in the other hand flux-schnell is harder to remove less quality but is apache 2.0 license. you are planning to test some of this models soon?
Cthey can't be undistilled. i've seen that claim posted as well. what he's doing isn't undistilling them, he's jsut breakign them worse then they already are
Dhttps://huggingface.co/ostris/OpenFLUX.1
https://huggingface.co/nyanko7/flux-dev-de-distill
this two projects looks promising and legit
https://huggingface.co/nyanko7/flux-dev-de-distill
this two projects looks promising and legit
Ci'm very aware of his projects. he is not undistilling anything
Che jsut thinks he is
Cwhat was done to flux can't be undone without getting the actual base model and retraining it
Cand he doesn't have access to that, or the funds to do that, or the machines needed to do that
Dok.. may be you right, I'm not an expert or a software engineer
Cthat doesn't mean you might not get good images with his code. you might - but it's not because he's undistilling it
Dsomething is doing, dome people claimed that the modified models can produce images without distilled cfg, and they work with a regular cfg scale
Dwhat I want is to train the models without bleeding and concept collapsing
Ci just spend a month walking through the flux dev and flux pro latent space, one token at a time. what was done to flux to hide the warping and shortening issues that everyone saw with the sd3-2b-medium release (because flux is the exact same neural network and setup) is really ugly, and undoable
Cthen train SDXL - flux is also untrainable.