Zellig AI content instructions
Zellig AI art
- Install Stable Diffusion ( or EasyDiffusion,but recommending SD automatic1111 webui, will be using it for the rest of this instruction).
- Find a base model, and most improtantly , get Ongezellig girls LORAs. (You can also try using VAE to improve the overall quality)
-Then drop the base model to stable-diffusion-webui\models\Stable-diffusion ,LORAs go to stable-diffusion-webui\models\Lora
(We will be using AbyssOrangeMix3 checkpoint as base, and Zellig LORAs can be found in here) - Generate some basic images with prompts. Use danbooru tags.For the example, Coco was '1girl, solo ,white ascot,yellow sweater, white skirt,tall, breasts, long hair, blonde hair, brown eyes,coco_ongezellig',
adjust the LORA strenth( <lora:name: value>, base value is 1, but suggesting using 0.7 or 0.8). Pick best one. - Open up your image editor (I use ms paint cuz I'm lazy) and make some alterations (such as removing some things in the background, draw a longer sleeves).
- Copy your image into img2img tab (inpainting).
- Add alter tags to reflect your changes .You can increase/decrease focus on tag with (tag:value), base value is also 1.0.
- Cover select areas with impainting brush. CHOOSE ONE FROM BELOW:
a) cover elements you want changed then select 'Inpaint masked' mask mode
b) cover elements you want unchanged (like face) and select 'Inpaint not masked' - Start generating images until you happy with current changes or you feel you're stuck with subpar results. Play with Denoising strength (how much of the original image is retained).
- Repeat steps 4-8 until you done .You can also copy the image to extras tab and upscale it if you think it looks like shit
- Post it on the Internet and make chuddies goon to it or something
example:
Zellig AI song covers
The voice models can be found in here, and we suggesting using Replay to make song covers , instuctions can be found in here