Dive Deeper into AI Art Creation: A ComfyUI Tutorial on Customization and Control
Вставка
- Опубліковано 7 чер 2024
- Tired of generic-looking AI art? Learn how to customize your ComfyUI creations and then take them further in Affinity Photo for truly unique, polished images. This video covers style transfer, style exploration, and fine-tuning controls. We also explore how to prepare creative, unique photo portrait ideas in Affinity Photo
#### Links from the Video ####
Get my Workflow here: / 102733765
IPAdapter Install guide: • IPadapter Version 2 - ...
#### Join and Support me ####
Buy me a Coffee: www.buymeacoffee.com/oliviotu...
Join my Facebook Group: / theairevolution
Joint my Discord Group: / discord
AI Newsletter: oliviotutorials.podia.com/new...
Support me on Patreon: / sarikas - Навчання та стиль
Great content. This is true AI assisted art. People saying that anything made with AI isn’t art are trying to suppress the artistic expression of others.
so so true, ... ( sampling in music had the same problem in the beginning - now almost everything is sampled )
I love content like this. Please make more!
This is the sort of thing I enjoy doing, you actually feel like YOU made something at the end.
Havent done any comfy for quite a while, but Im getting a pen and tablet next week and Ill be all in!
Excited to play with this in Krita now the plugin got style transfer 2 days ago.
Thanx 4 this inspirations, best wishes
This is the Quallity Content i am here for!. ANOTHER! *Thor smashes cup on flor*
Good job friend
I was in love with the idea of ComfyUI at first, then hated it when I tried it.
Then when I tested the speed of ComfyUI vs Automatic1111, I went with ComfyUI for 99% of my generating.
I have Forge and is supposed to be the same or similar to ComfyUI that looks like automatic1111, but, I don't know.
I still use ComfyUI 99%+ of the time, unless I can't do what I want.
So there are style loras out there that help the AI with style, but does the IP Adapter potentially make them redundant if you can feed in the style you want?
If so, this would save a ton of disk space...
Can u please provide alternative links to get workflow? Cuz patreon blocked in some countries(
I will try to look into that. can't you use a VPN?
@@OlivioSarikas all VPN services is blocked - only Tor, but Tor blocking connection too, cuz Patreon collab with Tor and don't allow to connect...
where to get the saicanny model?
can u please make video on "ELLA - A Powerful Adapter"
Love it. When I am looking for inspiration, I often go to MidJourney, write only a few words, but with maximum stylization, weird, and chaos parameters. :) If I'm lucky, I will get something that intrigues me enough to work in that specific direction for a moment. :)
Also, combining MJ and Comfy UI... As soon as MJ opens their API, I hope for a Comfy UI MJ node to further integrate the workflows. :)
I just love what we can do today...
Hi Olivio, afayk is it possible at the current state of AI to put true element A into true scenario B and keep the scene realistic and genuine to the average viewer, but in a (close to) seamless way? Like, to let AI take care of all the technical fuss like shadows, relighting, contrast, sharpness, saturation, color palette and whatnot, without knowing (almost) anything about traditional photo editing? In short, my goal is to get a "believable", "passable" result without altering too much the true identity of the source material, I want to be in the Oval Office lol :D
The difficulty I have with Comfy UI is its interface. I get hung up on the simplest of things, namely I cant keep a.normal ciew of all the nodes. The moment i touch my scroll wheel i zoom in and cant get back to a normal viewing mode. The nodes also is visually confusing, like looking at wires in the computer, which gets in the way of me looking at the controls to create the image. And the image size is too small, but Im sure it coukd be adjusted. Im not a tech idiot, but this interface is daunting, not inviting for a user to explore, much less create.
Anyone tried ChatGPT to spit out ComfyUI workflows yet? 😁
Would have taken me a lot less time for my profile pic (which is a part of a larger piece) if I had done a collage this way! Probably would have looked better too! 😂😢
Love me some comfyui though! It makes so much sense to me because I formerly worked in audio- a big piece of it is signal flow
Too much nodes. Too complex.
thats not art its engineering
You say that the AI process in this program 2024 is engineering but it will turn into more art and less engineering in time as people begin to understand it more and less geeky methods are introduced, just like paint mixing used to be chemistry before squeeze tubes containing paint made the process more artistic. A similar evol. occurred in photography so that, now, it’s art. But marble and clay sculpting has generally remained the same.
The first 3D animation made on a computer was plotted with a number graph and moved by figuring it out on a slide rule
say that to Nuke, Blender, and Pixar Renderman users 🤣
The prompt that comes out of our mind and turns to an image is an art indeed, so yes, it is ART
Engineering is an art in itself.
AI isnt art, not yet at least while impressive quality ive yet to see anything that has more depth than a puddle. makes pretty girls but they are always selfies looking at camera, how interesting
Listen to him ya'll (ComfyUI) I drug my feet for months before taking the plunge. Spent a month and 1/2 neck deep learning Comfy by-way-of UA-cam University. When the clouds finally cleared, and the sun came out my Ai videos are SO much better than when I was using WebUi.
You are not an artist and you are not creative.