Easy Background Cropping with Stable Diffusion 2.0 and AUTOMATIC1111
HTML-код
- Опубликовано: 5 дек 2022
- We walk through how to create a photorealistic image with stable diffusion 2.0 and remove the image background with the new automatic1111 depth mask cropping extension.
Discord: / discord
======= Links =======
URL for Extension: github.com/Extraltodeus/depth...
======= Misc =======
Negative prompt used:
disfigured, kitsch, ugly, oversaturated, grain, low-res, Deformed, blurry, bad anatomy, disfigured, poorly drawn face, mutation, mutated, extra limb, ugly, poorly drawn hands, missing limb, blurry, floating limbs, disconnected limbs, malformed hands, blur, out of focus, long neck, long body, ugly, disgusting, poorly drawn, childish, mutilated, , mangled, old, surreal
======= Music =======
Music from freetousemusic.com
‘Biscuit’ by ‘LuKremBo’: • (no copyright music) l...
‘Animal Friends’ by LuKremBo: • lukrembo - animal frie...
#stablediffusion #aiart #tutorials #techtutorials #automatic1111 
Наука
Man, you're up too! You're a very smart guy, and making these vids reinforces it. I've learned a bunch from you. Very cool! Now we need to make Avacado butter out of the chair next. 😀 Thank you for making these.
im an OZZZY, we occupy a highly gakked area of space-time
@@lewingtonn say hi to marty n doc for us will ya?! ;)
Great video! Really enjoy the way you explain new features!
Relate max, Kain!
Dude, you're awesome! Thanks for the coding and vids :)
Sick stuff! depth mask cropping is inspired.
I try to reduce the multiple panels by using the negative prompt "collage". I have the feeling it does reduce them, but it does not eliminate them. If and how much it helps still needs to be tested.
This is also super helpful for sending straight to the img2img in-painting tab
negative prompt. split image collage
Hi! Nice video, thanks! Can you tell me if there are any ideas? I generate a picture, then through "inpaint + fill + only masked" add a new object, is there any way to quickly and accurately pull the same object in a separate picture on a white background ? The main idea, to adapt SD to generate game art for the hidden objects genre.
You can Color Range out the foreground in any image or video editing application. What is the fuzz about white not being exactly at the foreground (which never was anyway)?
Hi there, lets say you have now cropped your avo archair, and want to keep the cropped chair, but change the entire background. How would you keep the cropped chair, and create an entirely new background?
I use Tencent ARC for background removal.
Hi, I want to start a profession in this line of work though I’m still learning I’m pretty good at it so far. I was wondering what kind PC settup should I get? Do I need more ram, better cpu, can I used an external gpu etc.
we did a video on GPU recommendations: ruclips.net/video/H-6DXU967bU/видео.html&ab_channel=koiboi and yeah, basically the GPU is the main thing you need to worry about for this.
Good luck!
Im trying to remove the background of an anime full body photo. Maybe ‘cause dont know how to. This tutorial is really good but for my drawing not working
Quick question! I can't use dpt_ large, it just freezes my SD, my graphics card is an RTX 2070 Super which i think you also have, so I'm curious how you got it working
the error i get is "UnboundLocalError: local variable 'model' referenced before assignment"
@@camprey unbound local error means there is a problem with the code, not your memory. Try updating the webui and the extension. I have the exact same card and can run dpt_large fine!
@@lewingtonn i updated both in order to use the new SD 2.1 but still cant use dpt_large 😢i guess thats just my luck lol
@@camprey still the same error? again "UnboundLocalError" means there is some kind of code problem... one last thing you can do is go to the automatic1111 github and search your error in the "issues" tab
Love your work Sir. Thank you. Side note: You speak really quickly at times and I find it difficult to follow and have to rewind. For your consideration. X
why do you have a suspicious amount of buff joe biden in your output folder
i.... no reason...
@@lewingtonn understandable, its not like i have 1289 pictures of donald trump and joe biden in a bathtub on my txt2img-images folder haha that would never happen right
@@reyhan0447 yeah ha ha ha, and if you DID have those images, there's no way you'd reach out to me on discord and sent them to me or anything like that ha ha ha
😢OpenCV(4.6.0) :-1: error: (-5:Bad argument) in function 'cvtColor' > Overload resolution failed: > - src data type = 17 is not supported > - Expected Ptr for argument 'src'
So I selected the extension, and activated it and since you were telling us how FAST it's going to be, and set a timer to show how FAST it will be, and I made my image I want to use, and I hit generate, all excited, and then you say 'oh it might take half an hour to download'.
Bruh
But in the time I spent whining about it, it's downloaded and I didn't know about this feature, so thanks :) I do like your videos and the way you explain a bit more than the other tutorials I've seen. It's nice to know what I'm actually doing, rather than changing a bunch of settings I don't understand.
There's a few different ways to remove backgrounds in the stock AUTOMATIC1111 interface. Is there a reason you downloaded an additional script? Are the stock functions not working? Just curious, because while I was previously able to change backgrounds now it seems I get a mixed bag of results, which by no means compare to your time of 11 minutes.. Would be interesting video or just a comment as to why this is the background method you're choosing. Thanks for the howto bro. Peace
error: FileNotFoundError: [Errno 2] No such file or directory: './externals/Next_ViT/classification/nextvit.py', how fix?
you have to run the install_next_vit .sh in the midas directory (which just does a "git clone" ). After doing that you'll encounter the same problem I did, which apparently has no solution
i have the same error
I get a TypeError: argument of type 'NoneType' is not iterable
launch module depends on flowdas-meta library and PYPI doesn't contain it. Where is everyone getting this lib ? You had to get it from somewhere to get this working.
it looks like flowdas-meta has been removed from pipy for some weird reason, I have no idea where else you would find it, sorry :(
can I say also: why the HECK are people removing depended-upon pipy packages???
@@lewingtonn well you had the extension working, so you must have it on your pc right ?
@@lewingtonn it's been there for so long, I don't think its disappearance is an accident
Why when i use dpt_ large i get an error!
what's the error?
@@lewingtonn probably same error as him... i get FileNotFoundError: [Errno 2] No such file or directory: './externals/Next_ViT/classification/nextvit.py'
@@alecubudulecu same here
Good videos, but pls! Raise or normalize your volume m8 :) Use a compressor, limiter or anything to get closer to 0db, this is like -15db
the penalty for giving audio advice on this channel is DEATH, get 'im lads!
@@lewingtonn 😅 Yelp!
love ur shirt mate, looks like p-jammies! ;)
it's so siiiiick!
👋
👋