>>42478987
1/2
so first of all, you need Stable Diffusion together with some UI. i would recommend Forge or AUTOMATIC1111 for beginners. ComfyUI is a bit overwhelming for beginners with all its graph nodes
Forge is the simplest to install as it has a one-click package that does everything so no need to use git, terminals and so on:
https://github.com/lllyasviel/stable-diffusion-webui-forge
once that is done, you need to download a proper model. you can find various models here:
https://civitai.com/
you can start with PDV6:
https://civitai.com/models/257749/pony-diffusion-v6-xl
though i'd recommend this one:
https://civitai.com/models/288584/autismmix-sdxl
it's based on PDV6 but with some other model/LoRAs mixed up. it's more consistent with hands and other things. pure PDV6 is also a bit outdated and V7 needs "two more weeks"
there's also autismmix-confetti which i've been using recently. even better consistency at the cost of other things
https://civitai.com/models/288584?modelVersionId=324524
once you install the WebUI and it opens a tab in your browser, you have to put the model file in */models/Stable-diffusion/, then in the freshly opened tab, chose the model with '1'. if it didn't show up, press button '2'. you're ready to gen!
'3' is where you put your positive prompts. (every prompt should be separated with a coma) this is where you tell AI what you want to see
while AI does understand phrases to some degree, it might struggle with more complicated ones so i usually divide everything into simple prompts. 4ex: you could write "blonde girl with blue eyes and green shirt" but you can do it like this: "girl, blonde hair, blue eyes, green shirt"
i've seen weight adding by adding exclamation marks, putting the prompt in brackets ( ) or by adding ':1.3' at the end of the prompt but in my opinion they're ordered first to last so if you want to put emphasis on some prompt, put it earlier on the list. there's also stuff like ADDCOL, ADDROW, BREAK etc for regional prompting but that's more advanced. baby steps
