- This topic has 8 replies, 4 voices, and was last updated 5 months, 2 weeks ago by
.
-
Posts
-
Now that the art competition is done and dusted, belated congratulations to our worthy winners. I thought I'd share some images Z image turbo spat out, that I'd considered, but decided not to post for various reasons.
I started off wondering, what would be the worst Winter imaginable? An old Tangerine Dream album, Springtime in Nagasaki, made me think it would be a nuclear winter. So the horrors of that are reflected in the drab clothing of the woman, and in the fact that even the first snowdrop, the first sign of hope for better times, can't bring a smile to her face.
Ultimately, I decided the theme was a bit too dark, so I opted for the brightly coloured toon style, which became my first entry in the AI section.
For my second entry, I wanted a hot spring theme and came up with this image.
The only thing I didn't like about this one was the splash; it just didn't look right, and while I was contemplating how best to remove it, I decided to go toon style for this one too, leading to the next image.
If it hadn't been for those oversized snowflakes in the foreground, this would have been my favourite of all the images created. I made several attempts to paint them out, but I wasn't happy with any of them, so I moved on until I arrived at the image that became my second entry.
Over on the DAZ side, I'd have liked to enter this image, but I'd already maxed out my entries, so I'll have to settle for posting it here.
So there they are, the ones that didn't quite make it.
Turbo is not trained for toons. I meant, toons is just couple percent of its training data. There are plenty of more suitable models for this.
@noa091999 I used a Pixar-style Lora to get the effect I wanted, and I was pretty happy with the result. If I hadn't got something I was happy with, I'd have switched over to a FLUX checkpoint, but for now at least, I prefer ZIT; it's quick and delivers reasonably consistent results. Here are a couple of other efforts that I considered but didn't go with.
I'm not really a cartoon fan, but I think it's interesting how the boots turned out. Overall, the renders actually looks really cute.
@lucafeelgood I churned out a lot of images as I tweaked the prompts to get to the image I was looking for. With the lone snowdrop, I got there pretty quickly. The hot spring was more problematic. I had images with the clothes and the boots, floating on the water. It took a while to get them lying by the side of the pool, and not in it. Throughout those iterations, though, the boots remained looking like... well, boots. That's one of the reasons I prefer using ZIT for my AI fix; it's consistent.
Incidentally, my 2 AI entries to the art competition only gathered 9 votes between them, so you're not the only one here who isn't a cartoon fan.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.
@legolas18 Looks good, I'll need to try that out.
Back to the boots. As I said, I think it’s interesting how the boots turned out. I actually wanted to say that the boots are quite big, but that’s what makes the render humorous. Really nice.
- You must be logged in to reply to this topic.
