#Is there a guide for installing checkpoints?
161 messages · Page 1 of 1 (latest)
hello! you need to put the ckpts in the models/stable-diffusion folder and then press the blue refresh button to get it to show up in the list 
hi hi
ok so
ive got the waifu diffusion main folder in there
is there a file within that folder that I need to pull out?
Hi
👋
If you're unfamiliar with Stable Diffusion you can learn it here
https://huggingface.co/hollowstrawberry/stable-diffusion-guide/blob/main/README.md#index
oh
alright so the guide has me download a safetensor file which makes sense
what if for the above case where the download gives me folder
is there a specific file i need to pull out of it?
im honestly not sure
let me take a look
n
no
doesnt seem like it
ahh I see
do i just delete this folder then?
yeah
those are for scraping the datasets used and trainer folder is for training a model I think
yeah
yup!
you can get the waifu diffusion checkpoints here: https://huggingface.co/waifu-diffusion/wd-1-5-beta2/tree/main/checkpoints
yeah i believe thats where i got it the first time
ok that explains a lot
do you personally have any tips for using automatic
and also, how do you do dark mode 😅
take it as you go!
just try prompting for now and generating images
im not sure cause i have a custom style in my webui, but there is a dark mode option in the settings I think
https://huggingface.co/Linaqruf/anything-v3.0/blob/main/anything-v3-fp16-pruned.safetensors this is a good beginners anime model to try as well
ooooh
custome styles
css im assuming?
yup!
hmm, is there somewhere I can get pull those from?
I just setup a quick css apply using the style here https://github.com/AUTOMATIC1111/stable-diffusion-webui/discussions/5813 This is a tool for designers to worry more about designing their css, ra...
this has some ones you can try
you can install it as an extension to the webui by going to the extensions tab -> install from url
then go to installed and apply and restart ui

you can try dark reader maybe
the demo image they have looks pretty good
but
...
is background supposed to be blankbackground.png?
hmm... I guess you need a custom one
strange
reinstall?
if you refresh it'll go back to normal
😭
sorry to keep asking
ive got the image browser extension here, but there arent any images from the few prompts ive tried
would you know how that works?
no don't worry! feel free to ask as much as you want, we are here to help you
hmm.. did you try pressing first page? sometimes it takes a bit for me to load the images
also you might need to restart your webui completely, sometimes the installed tab doesn't restart it correctly
itnerensting
ive also got this weird mosaic stuff happening on the images
would you know whats going on there?
hmm
what kind of mosaic stuff
oh I see, are you using the latent upscaler?
no latent is very good, but it can have that mosaic issue with loras sometimes if you are using any
is there anyway to control colour of the image?
the colour is sometimes very vibrant or very washed
I think there's a built-in darkmode as well
can you try appending /?__theme=dark to your url?
if it's local it should look something like http://127.0.0.1:7860/?__theme=dark
also the command line option --theme dark should work
thanks ao i couldnt figure out how to force it 
using a VAE can help with that. which model are you using right now?
im using orange mix
following the hugging face guide
yeah I did
ive also got dreamshaper
which says its got a built in VAE
im using other peoples prompts with the same seed but i cant seem to get anything similar
also how are these guys getting landscape/portrait resolutions?
are you using the same negative prompts and such as well?
usually people generate in 512x768 or 768x512
yep
so the one im looking at
modelshoot style, (extremely detailed CG unity 8k wallpaper), full shot body photo of the most beautiful artwork in the world, medieval armor, professional majestic oil painting by Ed Blinkey, Atey Ghailan, Studio Ghibli, by Jeremy Mann, Greg Manchess, Antonio Moro, trending on ArtStation, trending on CGSociety, Intricate, High Detail, Sharp focus, dramatic, photorealistic painting art by midjourney and greg rutkowski
canvas frame, cartoon, 3d, ((disfigured)), ((bad art)), ((deformed)),((extra limbs)),((close up)),((b&w)), wierd colors, blurry, (((duplicate))), ((morbid)), ((mutilated)), [out of frame], extra fingers, mutated hands, ((poorly drawn hands)), ((poorly drawn face)), (((mutation))), (((deformed))), ((ugly)), blurry, ((bad anatomy)), (((bad proportions))), ((extra limbs)), cloned face, (((disfigured))), out of frame, ugly, extra limbs, (bad anatomy), gross proportions, (malformed limbs), ((missing arms)), ((missing legs)), (((extra arms))), (((extra legs))), mutated hands, (fused fingers), (too many fingers), (((long neck))), Photoshop, video game, ugly, tiling, poorly drawn hands, poorly drawn feet, poorly drawn face, out of frame, mutation, mutated, extra limbs, extra legs, extra arms, disfigured, deformed, cross-eye, body out of frame, blurry, bad art, bad anatomy, 3d render
how do I get that resolution
landscape
768?
I think 512x960 then upscale
one other thing
ive tried to use character specific prompts but they always return very different results
exusiai \(arknights\)
doesnt really give me anything related
this is 960x512
models by default recognize only a handful of very popular characters
like hatsune miku, touhou girls and such
you might need to find a lora for exusiai
there's 2.5kish pictures of her on danbooru but that pales in comparison to miku or reimu
ah
i understand
are there any specific things i need to do with a lora
or just the installation and using it in the prompt
well you should be able to just drop them in the folder, then you press
1 the button
2 the lora tab
3 the lora you want
https://a.yuri.fun/nmbmq.png
and it will be added to your prompt
it's best practice for /creating/ lora to still tag everything, so you might need to use proper tags in your prompt, that should be explained on the specific lora page
on civit or wherever
Try pressing the refresh button on the right?
cant say i see anything
other thing is, if i download a checkpoint do i need to download a VAE for it as well?
the checkpoint should work without a vae; but for better detail (and i think preformance) you would need the vae
ah understood.
Another thing is, is there a prompt I can use for full body image
also, I have a 3090, but im avereage around 1.3 -> 4 it/s
reddit says im supposed to reclone the ui?
wait
maybe because im upscaling

https://a.yuri.fun/xujgk.png I mean this button, it seems the loras are in the correct directory so pressing that button should make them show up
yep that was it 👍
