The model comes in fp16 format and as a checkpoint - that means that all necessary components namely VAE, Clip and textencoder are included. So be sure to load the model as checkpoint. When loaded as diffuser model you will have to load VAE and textencoders as separat files.

As a base model this checkpoint can create all kind of images that you like: Art, photography and anime and many more.

After testing this model with different samplers and schedulers I really want you to use:

Sampler: euler_ancestral

Scheduler: beta

Steps: 8 to 12

CFG: 1.0, Distilled CFG: 3.5

A lot of other sampler, scheduler combinations work well. For higher sharpness use 12 steps.

If you are like me and like your images with a nice softness and done in short generation times you are allready fine with 8 steps.