Forum Replies Created
-
Posts
-
@felicia Thank you Felicia. Well, this is a new model (in the last year the amount of new models has been insane), and particularly good. The fact that it can also create audio is a huge plus too.
I am not posting much as I am trying to stay away from the PC during the Summer. Stay frosty 🙂
RIP AnjaBear. She was always great to talk to and helpful. I hoped she was getting better. Sending my heartfelt condolences.
Smart move to go from some great DAZ image and then using AI for animation.
Lovely job. I hope the PC didn't fry you too much 🙂
I love the pixar style images, they look really good and more consistent in style (due to the lora). 🙂
Lately, I have been testing the new Anima (currently reached preview 3, so it's not final yet) with Gothic Niji Journey lora (I just love that syle).
It is quite fast and small, and gives some really good results, but it's more character-centric and basically an upgrade to SDXL/Pony/Illustrious.
@d0gg0d Just busy with work, and maybe more absent minded than usual 🙂
I am glad you liked it, I thought it came out pretty well too. 🙂
@felicia I am actually using a checkpoint for Z-Image-Turbo called Moody Real Mix V5 DPO with a LORA called Simple tfbs (body weight slider) (which I set at -1.75) as well as some custom workflow I made 🙂So, it's all local.
Happy Easter!
@littleevie Thank you for the kind word, Eve 🙂
I am glad you liked it. Hopefully, I'll have a better schedule for the next one.
@inversekinematics I am using Win 11 with a couple of 2TB SSD. Nothing too fancy on that regard.
I render my AI images with comfyui (custom workflow).
@felicia Thanks Felicia, for now I am playing around with two new generators (Qwen and ZIT). As for video, I stepped back a bit because I was not getting the results that I wanted but, since then, new stuff has come out that I will eventually try out.
Thanks for enjoying 🙂
@felicia I haven't done any videos in a while, so I'll link a slightly older one. Nowadays there should be better tools (this one, infinitalk, seems to miss things sometimes), but I haven't tried them yet.
Vo Suna HD for Genesis 9:
AI:
Arguably using the AI to pick from others work means there are hundreds or even thousands, maybe millions of others involved in one image.That's like saying that commissioning from one artist is like commissioning from thousands. Art wasn't born in a day, but it is the cumulative work of thousands of people too. If it applies to AI, it applies to an artist too.
Yes I read it. a ton of prompts and a map to roll dice on. I acknowledged it took effort to create. Its easier to re-click a template than retype it over and over then fill in the blanks. I guess I could say that the effort I put behind responding with these words I'm typing now is art of sorts then.You are the one equating art to effort, not me.
Id say theres a fine line between inspiration and plagiarism. Lucas said there were several movies that inspired Star Wars. Perhaps he stole a bit from each but Id suggest that he changed it enough to make it his own. AI didn't write it for him.An artist and an AI both have the ability to plagiarize, but that doesn't mean that just because they are capable of something, they actually do it.
Seeing the flood of AI images produced by anyone with a computer, It somewhat supports my perspective.Seeing the low quality images produced by so called "regular" artists supports mine.
So you get AI to tell you what to type into AI? I'm probably missing something because Im not on the high IQ list and haven't used chatgpt but that doesn't sound very creative.That's because you use chat gpt to enhance the prompt, rather to invent it. ChatGTP tends to always make similar things. Either way, I enjoy creating art as an AI artist. Whether or not you want to acknowledge me as an artist or not, is completely irrelevant. Without me, those images wouldn't exist. AI doesn't do them on its own, and I wouldn't be able to do them without AI. Just like a wood carver wouldn't be able to carve wood with his bare hands. T
@d0gg0d Commission something means there are two parties. Making an image with AI on a local computer means there is only one party. So the comparison doesn't stand in that case (although, it does if you do it on a paid website).
I'll agree it takes time and some effort to type and constantly roll dice.You haven't even read my first message. I showed you the work required. What you are saying doesn't apply for local generation at all.
out a description then include tags like Giger, Cyberpunk then take credit for effectively plagiarizing anothers work.By that logic, everything is plagiarized, even other people's work, since they plagiarized it from someone else.
believing she isn't intellectual because shes talented is an interesting perspective.Being able to draw, doesn't make you an intellectual. If you want to draw like that you need a lot of muscle memory and a lot of training, as well a love for the art to spend such huge amount of time training. Not a superior IQ.
@d0gg0d
You tell all your friends how hard you worked to get the yard just right.That's like saying that an architect does nothing, because someone else made the manual labor. I think you are overvaluing the manual labor. I prefer the intellectual one.
@d0gg0d The "prompt hero" concept is archaic at this point. ChatGPT creates my prompts, not me. I just tell it to give me 10 prompts, indicating style, mood, etc, making modifications as I go.
The effort, as in anything in life, is only done by a few creative people, despite the inherent limitations or options of the medium.
In the case of AI image generation, that is achieved locally, and only by some people, creating a custom program (called a workflow in the lingo of a certain UI called comfyui where that is possible) to generate and process the image in multiple iterations/steps and testing and finding the right combination among the dozens of dials, options, nodes, etc. And that has to be done for each model (there are currently 4 main models I am using atm)
Workflow Example (one of many I have created):
These are the options in just one node:
So, I hope you understand now why i find your argument reductive. It applies to most people, but not to everyone. Also, it does take effort and ingenuity, just not an artistic one, but a technical one, and the prompt is just one small cog in the equation.
@hobbyist Longer than 5seconds is possible, but not by much, and at the cost of extra computational time and loss of coherence. It is easier and better to create shorter videos with an input image several times and then stitch the videos together.
And yes, it outputs all the frames as pics if you want, which is very helpful if you are using photoshop, since you can create videos with those.
