Really love how far Invoke has come. Such an awesome experience. Still would love to see lasso selection tool, clone tool in the canvas and flip/flop layers. Keep up these shorts and the livestreams, it's so informative!
Following the steps one by one I see that I don't have the same controlnet models under sdxl I don't have contour detection scribble I downloaded all the controlnet models from invoke and added others but they appear under sdxl t2i adapter
I do like Invoke, but that house in the desert really didn't work. I think that's a good thing though and shows that there is still a good long way to go! Good to see the bad along with the good :)
We definitely like to share both the wins and the struggles in these live studio sessions to show that AI isn't some kind of magic button. There's a *lot* of trial and error and sometimes, things just don't really work out.
It is VERY difficult to follow your cursor/mouse pointer, let alone understand what it did when it got somewhere. If you went half as fast, it would cut down on the stop-replay-stop-replay stop again cycle for us. Also, if we can see the end result first, each step will make a whole lot more sense. "Tell 'em what you're gonna tell 'em, tell 'em, then tell 'em what you told 'em. Basic teaching technique.
Come join our live stream on Discord! Much easier to follow along to a live stream live rather than after the fact :) discord.com/events/1020123559063990373/1332445691196342403/1334946299904000000
100% agreed! We'd love to have some professional artists come livestream with us but unfortunately most aren't willing because of fear of backlash. So until then, we'll stumble through to try to show what's possible.
Just looked at the pricing. Holy crap you charge $10 per hour to use this. It's just a bunch of free tools. Even Adobe don't charge that much for their entire suite, and before people call me cheap I'm subscribed to hundreds of pounds worth of AI tools a month so I do see value in paying when it's justified.
@@invokeai Hey. Thanks, I've been giving it a test run but already uninstalled it. The UI is dreadful and it's been made in a way where you have to jump through hoops to get additional files like clip, and encoders to install on flux models. The advanced window is greyed out. So I had to go on a wild goose chase to try and find where the files go in invoke (Something that Forge and Comfy make super simple), and then the image generation window greyed itself out for no reason that required a restart. Thanks but until that horrible UI has been sorted and things made way simpler to drop a model and it's requirements in are done I'll avoid.
Really love how far Invoke has come. Such an awesome experience. Still would love to see lasso selection tool, clone tool in the canvas and flip/flop layers.
Keep up these shorts and the livestreams, it's so informative!
Following the steps one by one I see that I don't have the same controlnet models under sdxl
I don't have contour detection scribble
I downloaded all the controlnet models from invoke and added others but they appear under sdxl t2i adapter
I do like Invoke, but that house in the desert really didn't work. I think that's a good thing though and shows that there is still a good long way to go! Good to see the bad along with the good :)
We definitely like to share both the wins and the struggles in these live studio sessions to show that AI isn't some kind of magic button. There's a *lot* of trial and error and sometimes, things just don't really work out.
Is this SDXL? Can u do this with Flux in Invoke?
This is using an SDXL finetune. Flux has a less mature control model ecosystem, but it is possible.
@ As good as SDXL is, it would be great of Flux could do this just as easily. Not much use for it if you can’t do basic things like img2img.
It is VERY difficult to follow your cursor/mouse pointer, let alone understand what it did when it got somewhere. If you went half as fast, it would cut down on the stop-replay-stop-replay stop again cycle for us. Also, if we can see the end result first, each step will make a whole lot more sense. "Tell 'em what you're gonna tell 'em, tell 'em, then tell 'em what you told 'em. Basic teaching technique.
Come join our live stream on Discord! Much easier to follow along to a live stream live rather than after the fact :) discord.com/events/1020123559063990373/1332445691196342403/1334946299904000000
what did we learn?
if you dont know basic perspective and design fundamentals, AI is not going to help.
100% agreed! We'd love to have some professional artists come livestream with us but unfortunately most aren't willing because of fear of backlash. So until then, we'll stumble through to try to show what's possible.
Just looked at the pricing. Holy crap you charge $10 per hour to use this. It's just a bunch of free tools. Even Adobe don't charge that much for their entire suite, and before people call me cheap I'm subscribed to hundreds of pounds worth of AI tools a month so I do see value in paying when it's justified.
We give this away for free open source. www.invoke.com/downloads
@@invokeai Hey. Thanks, I've been giving it a test run but already uninstalled it. The UI is dreadful and it's been made in a way where you have to jump through hoops to get additional files like clip, and encoders to install on flux models. The advanced window is greyed out. So I had to go on a wild goose chase to try and find where the files go in invoke (Something that Forge and Comfy make super simple), and then the image generation window greyed itself out for no reason that required a restart. Thanks but until that horrible UI has been sorted and things made way simpler to drop a model and it's requirements in are done I'll avoid.
You should apologize. They literally give everything for free and post a ton of free videos that explain how to use their software.
@@TPCDAZMaybe you could try to read the manuals or watch the tutorials to learn tools, instead of complaining so much ;)
@@JVillalbaAnimation and what is there to learn about a terrible ui?. Not very bright are you.