Pencoded TheLustriVA

DD 5.4 settings | Snow-covered village | Setup 1

This batch is the prototype batch for the artwork in an upcoming super-collaboration on YouTube: Directed by Eastgate Audios, written by LupinStoleMyHeart, and produced by TheLustriVA.

Example output

Leti's Planets Reference

An excellent scene.

The Result

Original Settings

These are the original settings I used for the batch.

Discoart Python Code

from discoart import create

da = create (

Attempt to generate a coherent character Part 2

I'm still looking for the right settings, and might begin with my previous attempts with a kitsune bard.

Of the three examples below, most break the ["description", "artist", "modifiers"] list convention, and all are relatively uncluttered descriptions.

Primary Settings

Setting	TheHistorian	WAS

Version 0.2.1 Released

There's a non-zero chance this dataset will become private in the next few days.

https://drive.google.com/drive/folders/1iPVp3TPSQ0PuPeXcpuBwm74yKB62rpXN?usp=sharing

A conversation with GPT-3 on the basics of training an AI

Even with some formal postgraduate education in machine-learning, lately I've been getting lost in the complexity of the subject's technical literature.

What follows is my first attempt at using Codex's code-explanation mode to drill down from some recent ML code to concepts I am more confident I have a solid grasp on.

I hope that this and any potential future 'conversations' illustrate that:

Everyone needs a refresher, especially for disciplines as complex as machine learning.

A conversation with GPT-3 on the meaning of "attention resolutions"

An acquaintence on Discord came across the term "attention resolutions" in the code for configuring the training of an image model and couldn't find documentation on the term's meaning. I could only find it mentioned in one paper and the way it was used didn't help.

The following 'conversation' came from having Codex explain a code snippet that likely exists nowhere else because I imagined it purely to try draw an explanation.

It seems to have worked.

	{
	"text_prompts": {
	"0": [
	"portrait of a beautiful fox woman with fox ears , gorgeous ginger hair with white highlights, green jade eyes, wearing leather armor, clear dark complexion, symmetrical face, Instagram aesthetic, cinematic, WLOP, artgerm, Jason Chan, Charlie Bowater"
	]
	},
	"image_prompts": {},
	"clip_guidance_scale": 5000,
	"tv_scale": 50000,
	"range_scale": 50000,

	{
	"text_prompts": {
	"0": {
	"0": [
	"A colorful portrait of a beautiful female athlete, by Alessio Albi and Nina Masic:3",
	"vertical portrait, center weighted, f1.8, 50mm, fujifilm, film grain, studio lighting, rim light, leica, hasselblad, colorful light leaks",
	"artstation,deviantart,instagram, hyperrealism:1.5",
	"glasses, signature, freckles, blemish, watermark, specular highlights, jpeg artifacts:-1"
	]
	},

	{
	"timezones": [
	"Africa/Abidjan",
	"Africa/Accra",
	"Africa/Addis_Ababa",
	"Africa/Algiers",
	"Africa/Asmara",
	"Africa/Asmera",
	"Africa/Bamako",
	"Africa/Bangui",