It’s hard to give precise figures, because there’s always tricks to getting a little more or less but from my (admittedly limited) testing SDXL is significantly more demanding, and 10+GB of VRAM is probably going to be the minimum to run it. I don’t remember exactly what I was doing but I run on an RTX A4500 card, and I managed to max out the 20GB of VRAM just with one SDXL process, where I can normally run a LORA training and 512x768 size images at the same time.
Protip - If an image is good but not quite perfect, stick to the same seed and use the X/Y script to run the image lots of times at different CFG levels.
A lot of the time I try to just let images come out as the AI imagines them - Just running img2img prompts, often in big batches, then picking the pictures that best reflect what I wanted.
But I do also have another process when I want something specific, which involves doing img2img to generate a pose and general composition, flipping that image into both a controlnet (for composition) and a segmentanything mask (for latent couple) and then respinning the same image with the same seed with those new constraints. When you run with the controlnet and the mask you can turn the CFG way down (3 or 4) but keep the coherence in the image so you get much more naturalistic outputs.
This is also a good way to work with LORAs that are either poorly made or don’t work well together - The initial output might look really burned, but when you have the composition locked in you can run the LORAs at much lower strength and with lower CFG so they sit together better.
The real value of SDXL isn’t the higher native resolution, its the improvements in rendering fingers and text and so on. But honestly I have not yet been super impressed by SDXL, in the same way that I want to stay playing the old game with all its DLC and mods. SDXL is good, but until we have the same depth of resources available I am staying with 1.5.
The community will decide what is best by which model they support
I am planning on cooking a LORA today - I’ll give this a go and report back.
Most SD stuff requires specific versions of everything, and as you say the documentation is poor even on Windows. Try other forks, and you may get lucky.
How is it meaningfully different to the existing Scribble and Lineart controlnets that are already working in Automatic1111?
Prompt was presumably “Shaq to the moon!”
You really think people would spend a lifetime writing books if they couldn’t make money from it?
Things which are free have no value, both economic and societal. Even when we pirate stuff, at least our society encourages creative labour.
Controlnet of some sort is your best bet. You’ll probably need to render an imperfect scene, spit it out as a lineart png, take to photoshop and draw the lines in more correctly, stick it back into controlnet (without pre-processor) and re-gen with the same seed and prompt.
You can do multiple rounds to get it just so.
You can just doodle a guide image for SD, but I am not good at it so I spin through until I find something I can fix.
Spotify sucks, but the whole music industry has sucked like that for literally a hundred years - A very very few artists make bank, about 5% make a little, everyone else makes zero.
There are lots of cars you can’t get parts for dude.
Wouldn’t it achieve more to boycott things instead? If you won’t even give up watching a tv show, you aren’t an activist you are just complaining on the internet.
He who pays the piper calls the tune. Don’t complain that modern media is garbage that doesn’t cater to you while also saying middle class soccer moms can sponsor everything.
Indeed. And that’s without considering that a lot of SaaS stuff on the consumer level lets you cancel at any time. Ok, you can get burned for 30 bucks if it turns out not to be all that useful, but the full packages are typically priced somewhere between eyewatering and “ARE YOU FUCKING KIDDING?”, and they always have been.
A perfect example here - GeForce Now costs like 20 per month, cancel whenever you like. A 4080 gpu costs way over a grand. It’s up to you whether you prefer to own, rent or not bother at all, but it doesn’t take a lot to convince me to spend 20 bucks, but it does take a lot to get me to stump up for a whole new PC.
It does matter though - The price paid to the creator was based on the prospect of X number of sales or Y numbers of adverts. Almost everyone who presently is trying to get their creative works seen is hoping that being seen helps them to “make it” and be able to write or sing or whatever as a full time job.
It’s true that SaaS does stop you from owning software… But what good does “owning” a piece of software do you if you can’t get updates anyway? Back in the pre-internet era we got used to software existing as discrete versions but it hasn’t been like that for a LONG time. As soon as patching became a regular occurrence, “ownership” became a service contract with a CD attached. Then the CD vanished, and it just became a service.
While I do dislike needless “as a service” stuff, that model does genuinely suit a lot of people. It’s not a conjob; companies offer this stuff because a lot of customers want it. Most of the companies that are selling you SaaS stuff themselves use SaaS things in-house.
It does seem to work fairly well, although I will say that it doesn’t fit my workflow at all so I haven’t done a lot of testing. I do think there are some UI things that you could look at though. Engine and Dimensions shouldn’t be minimizable lists, because the fields only take up as much space as the label does. Also, your tooltips are outrageously large, covering about 75% the width of a 1080p monitor which makes them quite hard to actually read.