Great information about camera angles. I appreciate you taking the time to explain these steps. This is a big part of my learning experience. Every day I learn something new with Stable Diffusion. Subbed.
me too actually! but it also means i need to merge it with every model i like to use it with, kind of a hassle... textual inversion is also an option but you can only use them in the same model you trained them in, and their quality is kind of worse than training a checkpoint.
there's a bunch of progress in that regard, instantID and Lora's for example. But yes, the artist would need to tweak it most of the time to get consistent results
@@RexelBartolome Tweak it, yeah more like rebuild It , because I have tried all of them. If you have an imported Image of a characters costume it will never reproduce it exactly . Lora's are only good for getting the characters likeness and body type down , clothing is all over the place . Plus Alot of apps have cencorship like , no fight scenes, no guns ; no celebrities ; and no bikinies
I'm generally against ai content. I'd like to learn how these things work but I don't want to support ai RUclipsrs by watching them, but seeing as you are an actual artist that actually paints yourself I appreciate your content and take your opinion on these things a lot more seriously.
lmao. Honestly, i know what you mean, I feel the same way when other youtubers clickbait AI titles about turning a profit, getting passive income, selling artworks etc. it's annoying to see. It's like they see this thing as another get rich quick scheme, but its what drives clicks and views so I guess I can't blame them either 😔
from my experience, "viewed from above" and "viewed from below" gives okay results in general as well. Thanks for the tips!
cool! didn't know that. thanks for sharing :)
birds eye view, awesome , thank you sir.
you're welcome :D
Great information about camera angles. I appreciate you taking the time to explain these steps. This is a big part of my learning experience. Every day I learn something new with Stable Diffusion. Subbed.
youre welcome!
that's absolutely fantastic, thanks so much for sharing, I was thinking of training a model specifically for camera angle and shot composition.
me too actually! but it also means i need to merge it with every model i like to use it with, kind of a hassle... textual inversion is also an option but you can only use them in the same model you trained them in, and their quality is kind of worse than training a checkpoint.
@@RexelBartolome LORA its insane !
@@lordkoke ohhh havent thought about LORA. But dont you have to merge it to a specific model so you can use it in A1111?
what is the prompt for a green/chromakey background?
Ai is crap for story boarding and making comics , The likenesses and clothing all ways changes from scene to scene
there's a bunch of progress in that regard, instantID and Lora's for example. But yes, the artist would need to tweak it most of the time to get consistent results
@@RexelBartolome Tweak it, yeah more like rebuild It , because I have tried all of them. If you have an imported Image of a characters costume it will never reproduce it exactly . Lora's are only good for getting the characters likeness and body type down , clothing is all over the place . Plus Alot of apps have cencorship like , no fight scenes, no guns ; no celebrities ; and no bikinies
I'm generally against ai content. I'd like to learn how these things work but I don't want to support ai RUclipsrs by watching them, but seeing as you are an actual artist that actually paints yourself I appreciate your content and take your opinion on these things a lot more seriously.
lmao. Honestly, i know what you mean, I feel the same way when other youtubers clickbait AI titles about turning a profit, getting passive income, selling artworks etc. it's annoying to see. It's like they see this thing as another get rich quick scheme, but its what drives clicks and views so I guess I can't blame them either 😔