Updated: Jul 16, 2024
base modelVerified: 5 months ago
SafeTensor
Type | |
Stats | 220 |
Reviews | (15) |
Published | Jul 15, 2024 |
Base Model | |
Hash | AutoV2 D541B72C49 |
I have made a lightning merge in the past, but by request, I have made an updated version and split it from my main model.
I think for fine details of lightning is not quite as good as my main SDXL model (which can do good images as low as 6 steps but not 4 step like this one.)
Settings I strongly recommended:
Sampler: DPM++ SDE
CFG: 1.5-2.5
Steps: 4-7
It should be good for people with slower hardware or just quickly exploring latent space.
Not good at more artful or anime styles but can get some intresting results if you try and use prompts for them anyway.