I've been training a new lora file for SDXL that's produced it's first few image sets. The lora is deliberately trained to create a heavily stylized image with a look that has high levels of detail with a hint of a 3D anime style. I know this one won't be for everyone if you prefer something targeting realism, but I do like what it makes
Here's a handful of samples from some of these image sets.
Nice job on the Lora training... I have been 'playing' WAN2.1, and I have trained a Lora for mud/quicksand on that... it produces reasonable results... and in not bad time either (depending on how impatient you want to be! lol ).
Love you, too
).
