daemonrat
almost 3 years ago

It's a merge. Creates some interesting results. Can be used at 512x512 and 768x768. Based on 1.5.

~It's much less capable at making MIA related images than the original model, but the art style remains strong.

Seems to work well with story type prompts. Example: You find yourself standing on the edge of a vast, rolling ocean, with waves crashing against the shore. The sky above is a deep shade of purple, with swirling clouds that seem to dance and twirl in the wind.

As you take in the breathtaking view, you notice a faint glow on the horizon. As you watch, the glow grows brighter and more intense, until it is almost blinding. Suddenly, a bolt of lightning splits the sky, and you feel a jolt of electricity pass through your body.

You realize that you are standing in the midst of a powerful storm, with winds howling and rain lashing against your skin. You feel a sense of awe and wonder wash over you as you watch the storm rage on.

But as the winds pick up and the waves grow higher, you also feel a sense of danger. You know that you must find shelter, or risk being swept away by the powerful forces of nature. Do you brave the storm and search for shelter, or do you turn back and seek safety? The choice is yours.

______________________________

Note from creator(s) of MIA:


I recommend using the following always at beginning of your positive prompt: masterpiece, best quality;

I recommend using the following always as negative prompt: lowres, bad anatomy, bad hands, text, error, missing fingers, extra digit, fewer digits, cropped, worst quality, low quality, normal quality, jpeg artifacts, signature, watermark, username, blurry

(These are NovelAI's prompt template settings)


PROMPT FORMAT: Character, tag1, tag2, etc; Character2, tag 1, tag2, etc; Location (this is the format the model is trained with, it's not necessary but recommended for best results)


explanation for this prompt format: Commas separate subtags for a major concept or character while semicolons act as major concept separators for generating multiple characters or marking the location at end of prompt.


EXAMPLES: 

masterpiece, best quality; Riko, shift dress, red whistle; orphanage

masterpiece, best quality; Riko, delver outfit, red whistle; Reg, aubade helmet, aubade cape, backpack; upper second layer

masterpiece, best quality; Riko, delver outfit, red whistle; Reg, aubade helmet, aubade cape, arm extended; lower second layer, seeker camp

masterpiece, best quality; Habo, black whistle; Reg, aubade cape; the first layer, eternal fortunes


If struggling to generate a character solo, try adding tags like "masterpiece, best quality, solo; Riko, shift dress, red whistle; orphanage"




CHARACTERS AND UNIQUE SUBTAGS:


Riko

shift dress (dress she wears in Orth)

cold outfit (she wears this in the fifth layer in the movie)


Reg

aubade helmet

aubade cape

arm extended


Shiggy (Shiggy had issues when captioning "red whistle" so I had to caption all of his images without it otherwise he would never get his red eye color. Generating him as a result will always try to generate something in place of his red whistle unless you add the red whistle tag for it to understand)


Nat


Kiyui

bell


Jiruo

child (placed before name "child Jiruo")

moon whistle


Habo

black whistle


Marulk

maid outfit


Ozen

ozen white whistle (you need to add this tag to generate her whistle)

ozen armor (for her armor outfit)


Lyza

lyza white whistle (to generate her whistle)

child (placed before name "child Lyza"


Bondrewd

bondrewd white whistle (to generate his whistle)


Nanachi

abyssal hat (Nanachi's hat)

young (placed before name "young Nanachi")


human Nanachi


Mitty


human Mitty




GENERIC CHARACTER TAGS: (unique ones are included since they may work on any character anyway)


shift dress (Riko's white dress she wears in Orth)

delver outfit (standard cave raider outfit)

delver helmet

aubade helmet (Reg's helmet)

aubade cape (Reg's cape) (Not advised to use on Riko, may result in distressing generations)

maid outfit (Marulk outfit)

bell

red whistle

blue whistle

moon whistle

black whistle

ozen white whistle 

lyza white whistle 

bondrewd white whistle

backpack

ozen armor

blaze reap

abyssal hat

bleeding curse (fourth layer curse)




LOCATIONS AND UNIQUE SUBTAGS


Orth town

the abyss

windmill

gondola platform

orphanage


orphanage


the first layer

gondola

windmill

eternal fortunes

cliffs (not an official tag in captions but generates cool outputs)


upper second layer (the forest of temptation)


lower second layer (the inverted forest)

seeker camp (for exterior shots of seeker camp)


seeker camp


the third layer


the fourth layer

garden of eternal fortunes


nanachi's hideout

exterior

backyard

eternal fortunes


the fifth layer


ido front

exterior

playroom




MONSTERS AND CREATURES


crimson splitjaw

corpse weeper

orb piercer

narehate

Meinya (You can sometimes generate characters holding Meinya with a subtag like "holding Meinya"

Read more...

What is MIAdream?

MIAdream is a highly specialized Image generation AI Model of type Safetensors / Checkpoint AI Model created by AI community user daemonrat. Derived from the powerful Stable Diffusion (SD 1.5) model, MIAdream has undergone an extensive fine-tuning process, leveraging the power of a dataset consisting of images generated by other AI models or user-contributed data. This fine-tuning process ensures that MIAdream is capable of generating images that are highly relevant to the specific use-cases it was designed for, such as anime, dreamlikeart, mia.

With a rating of 0 and over 0 ratings, MIAdream is a popular choice among users for generating high-quality images from text prompts.

Can I download MIAdream?

Yes! You can download the latest version of MIAdream from here.

How to use MIAdream?

To use MIAdream, download the model checkpoint file and set up an UI for running Stable Diffusion models (for example, AUTOMATIC1111). Then, provide the model with a detailed text prompt to generate an image. Experiment with different prompts and settings to achieve the desired results. If this sounds a bit complicated, check out our initial guide to Stable Diffusion – it might be of help. And if you really want to dive deep into AI image generation and understand how set up AUTOMATIC1111 to use Safetensors / Checkpoint AI Models like MIAdream, check out our crash course in AI image generation.