I always do the images on site because i love that people would achieve the same quality without paying for a GPU or something like that

The last version (v4 is like the official one, if there is a way to call it), trying to aim to be one of the best loras out there and that is why i change the name and trigger word to brand it properly. the first 3 versions are tests but looks good

Just playing around with training, just a realism lora (The last versions are the same, i just update some captions and to test something else here on civitai, obviously i'm not goona tell you what i'm testing hehe 😘)

Enjoy 😘 This my Patreon https://www.patreon.com/natmontero96 , if you want to support me, also you can send me some tips on https://ko-fi.com/NatMontero96

Follow me on twitter 😘 https://twitter.com/NatMonteroC

Comissions Open, DM me

Works good on 1 Strength, works better if you use up to 1 Good quality without upscale and high res, but obviously looks so much better with the high res and upscale

Sampler (Recommendation):
if you use just one it works good with:
Euler - Simple

Er_sde - beta

2 samplers:

qin_zhang_2s/diag implicit and qin_zhang_2s/diag implicit

qin_zhang_2s/diag implicit and deis_2m/multistep

Scheduler Beta/beta57

ETA .5

V4 Trigger word: Natty

Trigger word: NattyK