- π£ INFO Notice
- π£ INFO Terminology
- π£ INFO Multiples of 64
- π£ INFO [WIP] Preset prompts and negatives
- π£ INFO [WIP] Samplers and Schedulers on INAI
- π£ INFO Interesting prompts
- π£ INFO Undesirable prompts
- π£ INFO [WIP] Artist names
- π£ INFO Character Design Prompts
- π£ INFO [WIP] Scoring AI generated images
- Bottom page
π£ INFO Notice
This article was last seriously updated in the last quarter of 2024, right before the release of NAI XL v-pred v1.0 (also Laxhar Lab's "last" version). Additionally, I'm on a long hiatus from image generation models due to my personal dissatisfaction of my findings. Today (whatever date it is right now), some information may be useless or outright wrong. Make of that what you will about the contents of this page.
π£ INFO Terminology
- INAI: IllustriousXL/NoobAI-XL abbreviation. Caveat: I mostly use NoobAI-XL now, so some techniques may not translate to IllustriousXL. I just assume it works.
π£ INFO Multiples of 64
64 | 128 | 192 | 256 | 320 | 384 | 448 | 512 | 576 | 640 |
---|---|---|---|---|---|---|---|---|---|
704 | 768 | 832 | 896 | 960 | 1024 | 1088 | 1152 | 1216 | 1280 |
1344 | 1408 | 1472 | 1536 | 1600 | 1664 | 1728 | 1792 | 1856 | 1920 |
π£ INFO [WIP] Preset prompts and negatives
INAI and its finetunes
Prompt: original, cover image, album cover, textless version, newest, best quality, very aesthetic, absurdres, masterpiece
Negative: (speech bubble, sketch, lowres, text, cropped, worst quality, low quality, normal quality, adversarial noise, artifacts, jpeg artifacts, gif artifacts, bad anatomy, bad hands, fewer digits, bad proportions, signature, watermark, username, blurry, revision, comic, manga, 4koma, multiple views, chart, unfinished, chromatic aberration, artistic error, scan, abstract, artist name, censored, resized, ai-generated, ai-assisted:1.3)
Pony DiffusionXL and its finetunes
Prompt: score_9, score_8_up, score_7_up, rating_questionable, source_anime
Negative: score_6, score_5, score_4, source_pony, source_cartoon, censor, watermark, blurry, lowres
π£ INFO [WIP] Samplers and Schedulers on INAI
These resources are useful:
- https://www.reddit.com/r/StableDiffusion/comments/16wykzy/ever_wondered_what_those_cryptic_sampler_names/
- https://civitai.com/articles/7484/understanding-stable-diffusion-samplers-beyond-image-comparisons
- https://www.reddit.com/r/StableDiffusion/comments/1dohy20/quick_overview_of_some_newish_stuff_in_comfyui/
- https://stable-diffusion-art.com/samplers/
- https://www.felixsanz.dev/articles/complete-guide-to-samplers-in-stable-diffusion
Terminology
- Ancestral: Ancestral Sampling is a discrete, simpler method focused on generating diversity in outputs by injecting noise in reverse steps. Adds controlled randomness at each step of the reverse diffusion process, mimicking a generative "ancestral" process where randomness represents uncertainty or diversity in the model's predictions. It is tied to discrete timesteps and operates within the standard diffusion framework.
- SDE: Stands for Stochastic Differential Equation, a DE where at least one term is a stochastic process. Theoretically rich, continuous-time approach in creating randomness by using the SDE to model the forward and reverse diffusion processes as continuous-time dynamics.
Some samplers are dropped entirely because of low confidence of its capabilities. I have separated samplers into the following separate classes:
- Deterministic Samplers:
- Euler
- Heun
- Heun++2
- LMS
- PLMS
- DPM Adaptive
- DPM2
- DPM++ 2M
- UniPC BH2
- iPNDM V
- DEIS
- DDIM
- DDPM (deterministic, but different)
- Higher order needs much higher steps and lower CFG. Stochastic Samplers:
- DPM++ SDE, DPM++ SDE GPU
- DPM++ 2M SDE, DPM++ 2M SDE GPU
- DPM++ 3M SDE, DPM++ 3M SDE GPU
- Ancestral Samplers:
- Euler A
- DPM2 A
- DPM++ 2S A
- Dropped Samplers:
These are the schedulers that are tested. Typings are made from a combination of the observations of the results and how it works:
- Karras (Type 1)
- Exponential (Type 1)
- Beta (Type 1)
- Simple (Type 2)
- Normal (Type 2)
- DDIM Uniform (Type 2)
- SGM Uniform (Type 2)
- Align Your Steps 10, AYS (Type 3)
- Align Your Steps Generate in Iterative Steps, AYS GITS (Type 3)
π£ INFO Interesting prompts
Color
limited palette
sets an ambiguous multi-color theme.color contrast
sets an ambiguous contrasting multi-color theme to emphasize objects/things. relates to complementary colors.analogous colors
are colors close to each other.monochrome
sets an ambiguous one-color theme.spot color
to color in parts of the image to emphasize objects/things.partially colored
to color in parts of the image incompletely/unfinished.high contrast
greyscale
for black, white and grey colors.COLOR theme
whereCOLOR
is any of these: white, grey, black, red, orange, brown, yellow, green, aqua, blue, purple, pinkcolorful
for emphasis on colors.muted color
darker low saturation colors.pale color
lighter low contrast colors.pastel colors
very light colors.dark
is a softer alternative toblack theme
.light
is an ambiguous alternative to add lighting. See instead the specific section later below.
Styles, art techniques and types
lineart
for mostly line-only without colors or shading.sketch
for rougherlineart
no lineart
for objects/things rendered without lines/flat color
for objects/things rendered without highlights, shadows, shading, gradients. no depth.ligne claire
for a variant offlat color
where the subject and background has little contrast rendered equally.blending
for a variant offlat color
where the objects and surroundings blend together.neon trim
,neon lights
,neon palette
holographic clothing
,iridescent
Prints
Composition, perspective
scenery, landscape, wide shot, atmospheric perspective
combo to make scenic images. best paired with outdoor scenes.split theme
group picture
,lineup
,queue
,ghost in the shell lineup
face-to-face
,back-to-back
,front-to-back
,side-by-side
taking picture
,selfie
,finger frame
border
,picture frame
Lighting, lights, particles
diffraction spikes
,glint
sparkle
lens flare
light particles
Formats and Compositions: Subimages
multiple views
very generic tag. Should be handled delicately by filtering out more specific tagsvariations
is the more conservative sister ofmultiple views
.collage
comparison
very generic tag. should be handled delicately before training.zoom layer
,projected inset
split screen
,column lineup
,column lineup background
reflective floor
,reflective surface
,reflective water
,moon reflection
x-ray
Formats and Compositions: User Interface, Diagrams, Graphs
Others
(ANY flowers:0.6)
to reduce flower takeover.dual persona
,multiple persona
,clone
,time paradox
,look-alike
,selfcest
,twins
,twincest
,narcissism
age comparison
mirror
color coordination
,color connection
,rainbow order
deep penetration
fourth wall
against glass
,against fourth wall
,breast press
,breast press
,pectorals on glass
,ass on glass
,ass press
,transparent seat
,penis on glass
,pussy on glass
,cheek on glass
,hand on glass
invisible
,invisible penis
,invisible man
,disembodied penis
,disembodied hand
,disembodied tongue
,disembodied foot
paid reward
painting (medium)
faux figurine
π£ INFO Undesirable prompts
Metadata
These are metadata tags on danbooru indicating unwanted or undesirable images:
revision
,duplicate
,pixel-perfect duplicate
concept art
,production art
,prototype design
,design speculation
would mislead the model for established characters and concepts. Should be handled delicately.
Text, Symbols
Text, logo, symbols that reside in the artwork:
title
,group name
lyrics
seal impression
watermark
,watermark grid
,sample watermark
,too many watermarks
,signature
,weibo watermark
logo
,artist logo
,weibo logo
,nasa logo
copyright notice
artist name
,twitter username
,pawoo username
,circle name
,company name
,copyright name
,song name
,character name
,initial
,creator name
completion time
,dated
,timestamp
web address
mirrored text
,drawing on fourth wall
,segment display
,seven-segment display
Formats and Compositions: Subimages
These artwork formats/compositions I have low confidence in when generating or training. This category involves "subimages" or "self-similar subimages" that reside in the artwork:
photo (medium)
recurring image
,recursion
,double exposure
,kirigami
juxtaposition
,performance juxtaposition
,cellphone photo
,video game
,eyewear view
,iei
through medium
,through page
,through mirror
,through screen
,through painting
reflection
,different reflection
,bad reflection
,reflection focus
,eye reflection
,looking at mirror
x-ray
,x-ray vision
,x-ray phone
,x-ray glasses
,cross-section
,revealing layer
,page tear
internal cumshot
,cervical penetration
,impregnation
,fertilization
,female fertilization
,sperm cell
,ovum
,uterus
Formats and Compositions: User Interface
These artwork formats/compositions I have low confidence in when generating or training. This category involves interface elements that reside in the artwork:
user interface
screenshot
,anime screenshot
,game screenshot
,loading screen
fake screenshot
,fake phone screenshot
fake video
,viewfinder
,facial recognition
,recording
,progress bar
,livestream
media player interface
,play button
,fake play button
,pause button
,rewind button
,fast forward button
,shuffle button
youtube logo
,loading icon
,buffering
text messaging
,sexting
,chat log
Formats and Compositions: Graphs, Charts, Sheets
These artwork formats/compositions I have low confidence in when generating or training. This category involves images containing graphs, charts, sheets and other arranged information:
multiple drawing challenge
,color wheel challenge
,followers favorite challenge
,shirt cut meme
,six fanart challenge
chart
,bust chart
,character chart
,costume chart
,expressions
,expression chart
,kiss chart
,height
,height mark
,height chart
,move chart
diagram
,character profile
,relationship graph
,stats
,family tree
,information sheet
,alphabet
,gauge
,urine meter
graph
,bar graph
,radar chart
,pie chart
,cartesian chart
blueprint (medium)
,blueprint
,layout plan
how to
,making-of
,reference sheet
,sprite sheet
,artist progress
equipment layout
,weapon focus
turnaround
,clothes on and off
age progression
creature and personification
,vehicle and personification
Formats and Compositions
These artwork formats/compositions I dislike or low confidence in when generating or training:
music video
variant set
,large variant set
should be handled delicately by suppressing (but not entirely removing) their variants for training.sprite
as in game art and sprites. should be handled delicately before training.color guide
should be handled delicately before training.looking at self
,different shadow
bubble filter
circle formation
,weapon formation
floorplan
,diorama
,dollhouse view
,isometric
- these consist of visually guiding charts and sheets:
π£ INFO [WIP] Artist names
Currently, most artists are tested using the workflow below (please open in comfyui) using noobai-XL-Vpred-0.65s-cyberfix:
Terminology
Each entry has the following information:
artist tag
as taught to INAI.- Username linked to the danbooru artist page
- ~123 artworks count on danbooru
- 2 integer likeness score between 1 and 3. Higher scores correlate to closer style emulation to the original artist. Lower scores correlate to either ineffective emulation or adaptation of the original artist.
Here are various tags :
- [Multi-formats]: This artist has undesirable or unsuitable artwork formats for teaching models their style.
- [Exploratory]: This artist does not produce stylistically consistent artworks.
- [Biased]: This artist biases heavily in making artworks around a (very) small number of concepts reducing effectiveness on different kinds of art.
UNKNOWN
Not (thoroughly) tested artists:
GOOD
INAI succeeded or made acceptable results in learning the style of these artists:
nyantcha
Nyan β
~2200
toosaka asagi
ι εγγγ ~1900
tomose shunsaku
γγ’γ»γ·γ₯γ³γ΅γ― ~1700
kaamin \(mariarose753\)
γ«γΌγγ³ ~1600
fujima takuya
θ€ηζε ~1600
yd \(orange maru\)
YD ~1600
k-suwabe
γ±γΌγΉγ―γ ~1500
agahari
γ’γ¬γγͺ ~1500
dishwasher1910
D.I.S.H ~1300
modare
γ’θͺ° ~820
pottsness
pottsness ~590
missile228
MISSILE228 ~530
araneesama
Araneesama ~500
gsusart
GSUS ~470
dino \(dinoartforame\)
DIno ~470
remsrar
REMSRAR ~460
koh \(minagi kou\)
ζ΅·εͺγ³γ¦γKOHγ ~440
liduke
Liduke(ζ₯ε) ~370
wlop
wlop ~360
wangxiii
BW-ring ~340
turisasu
γγͺγ΅γΉ ~310 [unsuitable images]
reoen
γγγγ ~310
hara kenshi
γ―γγγγ ~300
ciloranko
CiloRanko ~200
z3zz4
Z3zz4 ~170 [unsuitable images]
namiki itsuki
εζ¨ ζ¨Ή ~140
BAD
INAI failed or made unacceptable results in learning the styles of these artists:
fune \(fune93ojj\)
γ΅γ ~110 [unsuitable images]
eita 789
γγγ ~90
eboda-x
ι’ θζ΅ζ°΄ε·₯ ~90 [unsuitable images]
yuyumu
ι±ΌδΏζ¨ ~67 [unsuitable images]
{30%:30%}
doo58455
DoodooπΎ ~45
{30%:30%}
π£ INFO Character Design Prompts
Each tag/prompt is sorted from the character's top to bottom. For example, hair details go first while lower body details go last.
Hatsune Miku
- V2 Design, default design:
vocaloid, hatsune miku, 1girl, aqua hair, hair between eyes, hair ornament, headset, twintails, very long hair, aqua eyes, aqua nails, aqua necktie, bare shoulders, black sleeves, detached sleeves, grey shirt, necktie, number tattoo, tattoo, shirt, sleeveless, small breasts, skirt, boots, thigh boots, thighhighs
- Append Design:
vocaloid, vocaloid append, hatsune miku (append), 1girl, blue hair, very long hair, long hair, hair ornament, grey eyes, grin, bare shoulders, high collar, sleeveless, sleeveless shirt, white shirt, alternate costume, black gloves, gloves, elbow gloves, bridal gauntlets, anklet, jewelry, blue nails, center opening, clothing cutout, navel, stomach, thigh cutout, thigh gap, black pants, no shoes
Kasane Teto
- Synth V Design:
synthesizer v, utau, kasane teto \(sv\), 1girl, ahoge, red hair, drill hair, twin drills, hair ribbon, hair bow, white bow, white ribbon, parted bangs, hair between eyes, hair ornament, shoulder boards, shoulder belt, black belt, flat chest, red trim, striped clothes, striped jacket, grey jacket, long sleeves, sleeves past elbows, petticoat, black sleeve cuffs, double-breasted, buttons, navel, grey skirt, striped skirt, layered skirt, skirt set, red belt, kneehighs, red socks, knee boots, lace-up boots, black footwear, cross-laced footwear
π£ INFO [WIP] Scoring AI generated images
1. Use an aesthetic scorer model
There are many aesthetic scorer models that aims to quickly determine whether an image is considered more "aesthetic" than another by reducing all of it into a single floating point number (the aesthetic score, ascore). Usually it ranges between 0 to 1, or 0 to 10. The current method used in INAI is waifu-scorer v3.
2. [WIP] Use an image noise estimation/measure
Find a technique that estimates image noise.
3. Manually measure properties of an image.
Depending on what kind of property you are searching for and differences of the image are, this method may not be effective way to determine an image's quality. Additionally, it will be hard to compare multiple similar images. I suggest you create a small score range for each thing you measure, like an integer between 1-3.
The following lists examples of things to look out when comparing images.
- Defects and quality of the anatomy, objects, things and background.
- Defects and quality of lines, gradients, patterns, colors, brush strokes.
- How things are composed.
- Presence of image noise. Use filters in an image editor like GIMP, Krita, Photopea or Photoshop to find them.
- Adherence to prompt. This includes artist styles, objects, poses, actions, etc.
- Flexibility in steps, CFG, samplers, schedulers, etc.
- Personal preferences.