Had an idea that would require learning how to utilize AI for text-to-speech projects, but I had to learn some things first. There's an Nvidia-specific application which sounded fun (Tachatron 2), but the RTX4090 is Ada Lovelace architecture and the app only supports Volta, Turing, and Ampere and the next one which looked interesting was ESPnet, ( Read more... )

To be completely fair, my life gets exceedingly complicated due to one variable and one variable only: Me. I'm the variable which complicates my life at seemingly each and every turn, without cessation. In this next episode of, "WTAF is Wrong with You?" I imagine at some point in the very near future, my wife coming across AI-rendered photos of ( Read more... )

• Portrait (2:3)
• 832x1248
• Standard (3:4)
• 880x1176
• Large Format (4:5)

Artificial Intelligence art-generators absolutely love booba; gigantic, over-the-top, gravity-defying tits. Even the most mundane, safe-for-work prompt will return whatever you've asked for...except with artifically-enhanced, award-winning chesticles prominently displayed just under tight-fitting clothes straining under the guise of the SFW keyword ( Read more... )

Time Machine AI model (LoRA) for Stable Diffusion; the classic Delorean from Back to the Future. You can download it here: https://civitai.com/models/118932/time-machine-back-to-the-future-delorean-dmc-12


( Read more... )

Re: Stable Diffusion/kohya_ss/ComfyUI

When you see this:

error: Your local changes to the following files would be overwritten by merge...
...Please commit your changes or stash them before you merge.
Aborting

Do this:

git fetch --all
git reset --hard
git pull

( Read more... )

Embedding Learning rate:

0.05:10, 0.02:20, 0.01:60, 0.005:200, 0.002:500, 0.001:3000, 0.0005

Two things happened simultaneously; my suddenly inadequate mid-range gaming rig due to the advent of AI, and my junior sysadmin-in-training wife deciding she wanted hands-on hardware experience after her initial CIS class. This was going to be a rather expensive, albeit unique, opportunity. Creating AI models from my photographs in order to ( Read more... )

The below is my configuration with RTX 4090

Dreambooth LoRA --> Configuarion file --> Open --> [confignew.json] --> Load

mkdir -p img/##_loraname [number of training images x folder repeats (##) x Epochs / batch size = < 3k]

Utilities --> WD14 Captioning --> Image folder to caption --> [00_loranameUtilities --> Blip Captioning --> Image folder ( Read more... )