Control-lora from StabilityAI + Revisions in ComfyUI
ฝัง
- เผยแพร่เมื่อ 4 พ.ย. 2024
- Want to control your Stable Diffusion AI art generations? Well, brand new and free from StabilityAI comes Control-lora! Ideal for use in ComfyUI with the provided example workflows and teeny-tiny controlnet files! 2.6 GB download? No thanks - I prefer 300 MB downloads :)
Enjoy!
== Links! ==
Install ComfyUI! - • Install Stable Diffusi...
ComfyUI Inpaint & Img2Img - • SDXL 1.0 - Img2Img & I...
ComfyUI ControlNet - • SDXL ControlNet Tutori...
StabilityAI control lora - huggingface.co...
ComfyUI Manager - github.com/ltd...
Nodes - github.com/Sta...
Installing Anaconda for MS Windows Beginners - • Anaconda - Python Inst...
== Even More Stable Diffusion! ==
Stable Diffusion Playlist! - th-cam.com/users/pl...
Add yourself with this Dreambooth Playlist - • Stable Diffusion Dream...
Textual Inversion Playlist - • Stable Diffusion Textu...
![[Part 1] ทารกน้อยถูกทิ้ง👶 หญิงชราเก็บขวดขายเลี้ยงดูเขาจนโต #โลกการละคร](http://i.ytimg.com/vi/nPTDvFCExks/mqdefault.jpg)
Revisions is my favorite so far - just dropping two random images in and seeing what comes out is SO FUN! I'm having a blast with it. Thank you for the great video, Nerdy Rodent!
Facts. Revision is crazy. The stuff coming out of it has me stuck 😶
Great timing, working on EBSynch, ControlNet workflow for videos
Perfect! Should be good 😀
Nerdy, gotta be honest... this is probably one of my favorite advancements thus far. Revision is INSANE. Thanks for being so quick on the tutorial for this. Pretty hard to go near auto1111 with all these comfyUI advancements. All we need is deforum at this point.
thank you for your awesome tutorials
Glad you like them!
Awesome, thank you
You're welcome! And thank you for watching :)
love your presentation here nerdy. I'm sold
Yay, thank you!
In Soviet AI-run world simulation, lora controls you...
Dude, I spent a good amount of time for having an automatic1111 that works smoothly.
Should I move to ComfyUI because I'm missing something, or are they pretty much the same?
They’re much the same, but ComfyUi is less rigid and allows you to basically made your own workflows. It’s also less of a resource hog and has support for the latest stuff “out of the box”, so to speak.
@@NerdyRodent
Thanks.
So the basic workflows come as presets, and you can manipulate those.
Cool, I thought I was supposed to create them all from scratch.
I'll give it a try for sure ;)
MANNAGGIA A TE CHRISTIANAIRTE!
ComfyUI is supposed to allow 4gb vram SDXL
@@NerdyRodent for your info Christian Ice is a VIP in Italy ;)
the timing on this guy : )
😉
Great tutorial one quick question though.... i have it working this way but what if I want to add one of my own LORAs on top of the control net. Like for example if I added myself as a character with a custom lora checkpoint. where do I put that in with all these nodes?
I added my own lora when trying this canny version, but it crashes after the 1st try. Works if i pick a different lora, but then it crashes on the next generation again.
very cool!
THANK YOU!!!
You're welcome, and thank you for watching!
Comfy may be a bit faster but it's a mess with noodles all over the place .Automatic 1111 has all the stuff that comfy doesn't and now it has the refiner extension for SDXL and a SDXL Styles extension to .Very cool .I spent a lot of time learning how to use Automatic 1111 and downloading and installing roop , controlnet and loads of other stuff.It's still my goto of choice .I like to use the seamless tile option and other stuff that comfyui just doesn't have yet .It's nice to be able to concentrate on creativity and artistic workflow rather than using pipe nodes and other nodes that promise a lot but disappoint .Maybe it might be worth using confyui in the future but not at the moment .It's still teething .
i wonder how revision works exactly? it seems to generate images with the same mood of the input, its very abstract and i love it a lot
Yup, it’s fun!
Cool
ImageScaleToTotalPixels node is not available or inaccessible from the manager. any idea why, and is there an alternative way to get it?
Btw, thanks for another great video...
I only have the extensions shown installed 😕
I have a problem, i follow all your steps in the video and try to blend two pics with the revision workflow. but it seems all my two images did not change anything with the outcome, the result is the same and it is not related to my images. Can I ask what is the reason of it?
Really sucks these aren't available for 1.5
Technology moves on.
Use SDXL 1.0 in its native 1024x1024 instead of 1.5s 512x512 and you will get fantastic results in automatic 1111
how can I train comfy UI to create a specific character I have and make them more real or more cartoon like
i am sorry can you please help me with something i have a text generation webui tab open but for some reason the ai isnt answering what could be the problem(who really appreciate any help)
You sound different somehow, do you have a cold or have you trained an AI voice?
I am still waiting for a good workflow for photobashing and editing *sigh*
Seems fine to me for that with overlay and stuff…
💛
👋💙
wanna bet nerdy rodent is skynet teaching us to use AI to learn human behavior
Lol 😂
What ????
You lost me at download comfy
the screen shot resolution is so low
Wait, control-loras is not really for Loras? It's just an alternative with more freedom to A1111? 🥲
Nope, control-lora is control-net but with lora 🙂