I've made a better AI video covering most of the same stuff but focusing on only one program. Check it out: "Become an AI Art legend with your iPhone, iPad or Mac " ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-cyFEvGJ0Khg.html
Awesome thank you. I ordered a Mac Studio and assumed it would be a long time until I could do AI generated images like I can do on my Nvidia PC. Neat ! 😁👍
Draw Things is surprisingly hard to learn, but it doesn't take long to learn. I've found results very promising once you learn how to use it. Beefy computer for sure, my M2 Max fans kick in after a minute and stay on until I stop generating.
Yeah, it’ll cause my M1 Max and my Mac Pro 2019 to rev up. I’d be realty curious how the M2 Ultra compares to my two Macs. Right now I don’t really notice a difference between either computer and the m2 iPad is about the same speed.
@@dmug ooh thanks for the reply man. I’ll definitely look into that. I don’t mind paying if I can get a product I can professionally use. Especially for using old/shitty archival footage for documentary work.
Great introduction - I am using first diffuser and am getting ERROR when I enter cat playing with ball of yarn and select generate. basically Eoorr generating NN outputs? I am trying on a MacBook Pro - not enough processor power?
As soon as I downloaded the app diffusers and tried to generate an image I got hit with the error Error Domain=com.apple.CoreML Code=0 "Error computing NN outputs." Userinfo=(NSLocalizedDescription=Error computing NN outputs.} If you could help me out on how to resolve this I would appreciate it
I've never seen this error but I can deduce that NN probably means that it's Neural Network, so probably trying to use the neural engine. I'm guessing you're trying to use the Neural Engine acceleration on an Intel Mac? I'd check to make sure, and set it to CPU/GPU only.
There’s too many variables to make just a blanket statement for speed because of the amount of steps, model used, renderer, resolution, LORAs, upscaling, whether opting to use ML cores which adds in an additional preprocessing for first run of a batch but then speeds up later images in a batch etc. 512 x 512 a like a minute or so at 25 steps but it can deviate a bit.
I made another about AI that might be of interest about Chat GPT and using 3rd party clients. ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-XGc3OHDqr5c.html