Consistency in Stable Diffusion | ControlNet Tutorial
Вставка
- Опубліковано 29 сер 2024
- #stablediffusion #aiart #animation
In this tutorial, we are going to learn how to create consistent characters using the control net extension for automatic 1111.
How to Install the Automatic 1111 Web UI:
• How to Install Automat...
Stable Diffusion Complete Beginners Tutorial:
• AI Generated Art - Com...
🎼🎼🎼🎼🎼🎼
Song: ÉWN - The Light [NCS Release]
Music provided by NoCopyrightSounds
Free Download/Stream: ncs.io/Reloaded
Watch: • ÉWN - The Light | Tra...
🎼🎼🎼🎼🎼🎼
💾💾💾💾💾💾💾💾💾💾💾💾💾💾💾💾💾💾💾💾💾
Between my full-time job and my family life,
I try to find free time to create content for this channel.
You can support me and help this channel keep growing:
paypal.me/Mike...
💾💾💾💾💾💾💾💾💾💾💾💾💾💾💾💾💾💾💾💾💾
💻 This Is My Development Setup (Affiliate): 💻
============
Main Monitor:
amzn.to/3M64qCJ
Secondary Monitor:
amzn.to/41Iu06A
Graphics Card:
amzn.to/3MpnXzd
CPU:
amzn.to/3I8nvCW
RAM:
amzn.to/42zqM6u
Keyboard:
amzn.to/3W5RFN4
Mouse:
amzn.to/3nTPcZs
Headphones:
amzn.to/3pz0By5
Microphone:
amzn.to/3OecJz3
=============
Tags for the Algorithm:
ControlNet Net
Consistent Characters
Install Automatic 1111
Deforum Extension
Stable Diffusion
AI Animation Video
Reference Only is an absolutely game changer, when coupled with Roop. Wow, Thank you bro 💲💲💲
This was a good tutorial. Thank you. Hope you make more.
Unfortunately it did not work for me. A completely different person was produced. I even followed your prompts. Other than that your tutorial was real good. Thanks
Please do more videos of your Unity 3D | Open-World Survival Game Tutorial Series!
Do not waste time for this method. Install ReActor plugin immediately and start loving your work. I really mean it. You only have to give it one reference image and it can generate other images in various angles and styles. It's insane what it can do comparing to ControlNet.
Thanks
Nice
IMO, control net's reference only doesn't work very well. It gives somewhat OK likeness or similar features, but usually a more generic and duller look. Better to train a LoRA or textual inversion for your character's face and body.
It did so slightly. Lora is for 1 piece right, while own created character can be used for lora? Do you know short good vid for textual inversion?
is this process works for our personal images also ? please help me sir
Does it only work with faces or can it be used with complete characters?
Would be nice to see how to create a intro/Cinematic cutscene with Stable Diffusion generated images.
this process not worked for my personal images , please help me sir, i given one my reference image , i given same prompts of u and same values of yours , but my results are not of my face , character totally changed , please help me how to get same charaters with different results like you , please help me sir
Reference-only doesn't seem to work for images created outside Stable Diffusion. There is a workaround that might work:
1) Put your image in the img2img Inpaint tab. We won't paint anything in, that way we'll get the exact same picture as a result.
2) In the resize mode, set the resolution to the same as your image. The fastest way for this is in the resize mode section: click on "resize by" and let the resize value to 1, it means it keeps the same resolution as the image.
3) Make a picture, you will get the exact same image except now it's been built in Stable Diffusion.
4) Set that "new" image in reference-only, and boom! You'll notice that it influences your results this time!
its not working at all
is this possible on ComfyUI as well?
i do the same thing, and i have a real good portrait from my character ( like your ref ppicture ) but FFS why the picture who is using the ref got an head like a ballon ? the head is always big as fuck and yes i use the same resolution like my ref picture has and on the other hand its blocked the whole thing to change clothes, the face reffernce dont work
👎
bla bla bla blaa