(nintendo and etc) directed by notthatclassic
Viewing sample resized to 41% of original (view original) Loading...
Description

Praying Ninetales! It's something I really wanted to make since I took a neat 3-week break to try out some limitations of some kinds. A bit personal, but...
I wonder what they're praying about

Metadata

(NOTE used img2img from a previous gen, see further down for details)
- Prompts: solo, feral, Ninetales, silky golden mane, furry chest tuft, fluffy, [several fluffy tails], orange accents, long eyelashes, fang, starry dreamy background, closed eyes, sitting, serene smile, praying, praying with one paw to chest, one paw on chest tuft, holding chest, content, comical heart, heart by paw
- Negative Prompts: duo, multiple characters, couple, faded colors, faded orange
- Settings: Model: knkLuminai_v10, Seed: 810501944, Steps: 34, CFG scale: 4.5, Original Size: 1280x1024, Sampler: euler_ancestral, Schedule type: normal
- Upscale: Denoising strength: 0.4, Upscale Factor: 2.0, Hires upscaler: RealESRGAN_x4plus
(^I use SwarmUI, and the upscale is technically called "refiner", but it's basically the same as Hi-res Fix, I believe.)

### For img2img, I used a previous generation for the image. Those settings are...
# Prompts: exact same as image above!
# Negative Prompts: exact same as image above!
# Settings: Seed: 810501944, CFG scale: 4.5 (all other settings are the same from the above, for model, steps, etc.)
# IMG2IMG-specific: Initial Image Creativity: 1.0, Initial Image Reset-to-norm: 0.2
(^I used SwarmUI, and it has those settings there^. Descriptions for those are...
Initial Image Creativity: Sometimes referred to as 'Denoising Strength' for 'img2img'. Higher values make the generation more creative, lower values follow the init image closer.
Initial Image Reset To Norm: This essentially lets you boost 'init image creativity' past 1.0. Merges the init image towards the latent norm.)

Inpaints

I tried my best to document all my inpainting below. In Krita, I used the same model knkLuminai_v10, with all inpaints using it's "Alternative" Quality setting (Euler a, Normal) with 59 steps and a CFG of 4.5.
> Chest Tuft Area: furry chest tuft, starry night sky in background, paw on chest, (feral paw), 4 digits
> Paws Below: (feral paws), 4 digits
> Paw On Chest: (feral paw), 4 digits, furry chest tuft below
> Head (just one pass over the head, not mane): Ninetales, furry, fluffy, long eyelashes, closed eyes, content, content smile, cute nose
> Mane Hair (after painting in orange tips): Ninetales, furry, fluffy, long silky mane, orange accents, vibrant colors, night sky behind
> Stomach Fix (first pass): underside fur, smooth fur, soft shadows
> Stomach Fix (second pass) + Leg Fix: underside fur, smooth fur, soft shadows, smooth blending, smooth colors, orange tail tip in foreground
> Tail Fix: smooth fur, soft shadows, smooth blending, smooth colors
> Other Leg's Line Fix: smooth fur, soft shadows, smooth blending, smooth colors

Also note: in case you don't know, I use Krita AI Diffusion to inpaint/edit, look it up if you don't know it! It's very convenient if you edit between a lot of your inpaints!

Blacklisted

    There are no visible comments.