SVNN001版本v1.0_Safetensors (ID: 399161)

This model is focused on on single character, Non-Nude close-ups.

Images without people are of exceptionally high quality too.
CFG is good between 4-10, and I use DPM++ 3M SDE Exponential @ 60 Steps for my images.

Nudes need to be 'dragged out of it', and if that's what you're after then the SVSN001 model is a better match.

Here are some comparisons:

This checkpoint favors closer in, more covered images compared to the SN model.

Similarly multiple-person images are possible, but not a 'feature' of the model.

描述:

This is identical to V1.0, but using the Safetensors format rather than checkpoint

训练词语: