This is a model trained off of images that I have taken both from illustrations in books I own, and ones I've found on the internet. It is a work in progress, and does have numerous issues, but I am quite happy with the state it is in now. It has come a long way since I started working on it several months ago.

It is not historically accurate. Do not assume that the generations you get from it are in line with what you would have seen in the pre-Cortez Aztec Empire. However, it IS much closer than anything you would get on any base-model.

Key Terms:

'ztc -- general trigger

tenochtitlan -- the capital city

Clothing terms (hit or miss, needs more work):

maxtlatl -- loincloth

chalcayatl -- net cape

xicolli -- jacket

tlatoani -- regal dress

ichcauhuipilli -- cotton armor

ehuatl -- armor

tlahuiztli -- armor

ixtle -- maguey-fiber

cueitl -- skirt

huipil -- sleeveless blouse

huipil -- blouse

quechquemitl -- ceremonial gown

tilmatl -- cape

tochomitl -- rabbit fur

cuechintli -- net cape

To lessen the 2D/illustrated look, put this in your negative prompt:

2d illustration art drawing sketch cartoon

Issues:

Wonky perspective in some generations -- cause: bad photographs of illustrations, can be fixed with proper scans

Vertical lines in some generations -- cause: creases in illustrations, can be fixed in Photoshop

Drab or offset colour tone -- cause: uncorrected colour in scans and photos

To-do list:

Use photos of actual extant Aztec cultural artifacts and ruins to further train the model

Use better scans/photos/copies of training illustrations

Fix some issues with captions (missing details with regards to the names of clothing)

Please do not sell images generated with this model, I did not create the training data nor do I own any of it.

-- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --

This is the first in a series of models I am planning on creating, depicting lost time-periods and cultures that are of immense value to humanity. I am hoping to create a model on ancient Mesopotamia, and am also working on a model based on the art of pre-Soviet Ukraine -- with the guidance and input of a Ukrainian art historian.

I've decided to stop uploading models of celebrities and singers that I adore. I'm disappointed that the AI community seems to really only want to use this incredible technology to create ????... and, well, non-consensual ????, even if fake, is pretty disturbing for me. I will leave the models I've already uploaded up, but from now on the only human figures I will upload will be fictional characters.