Illustrious:
A lora model of bentodraws, trained over 64 images for 1024 using NoobAI Epsilon-pred 1.1 as a training model
Trigger word: benrawtod
Weight: 1 is perfect (you can go up to 1.5 and as low as 0.8)
Pony:
trained over 64 images at 1024 using Pony Diffusion V6 XL as a training model
Trigger word: benrawtod
Weight: Please also refer to the Mirko images towards the end for more reference
0.8 -1 you get the the full strength also adding "simple background" tag will enhance the effect but it's not necessary
0.6 - 0.7 is what I would recommend you start with it's the sweet spot
0.5 you get a hint of the style of more so of the checkpoint you're using
SD 1.5:
Trained over 64 images for 768 using Animefull-final-pruned as a model
Trigger word: bentodraws
Weight: 0.7 is a happy medium between the lora and checkpoint, 0.8 and above you get more of the lora and less of the checkpoint
Additional information:
You can see the models I have used/recommend below
VAE I use
For anime I use R-ESRGAN 4x+ Anime6B upscaler
For realistic I use 4x NMKD Superscale upscaler
For semi-realistic I use 4x-Ultrasharp upscaler
For Sampling steps I go 40-50
For Hires steps I go 10-15
For Denoising strength I go 0.4-0.6
I use 3-9 CFG Scale for my pics
If you wanna support me ko-fi