great model. does exactly what it says on the tin, and seems to have minimal interference with the rest of the composition. does better with trigger words included in prompt.

Unfortunately, this LoRA seems heavily overtrained. At 1.0 strength it changes almost the entire image and strongly recreates the example images instead of just applying the low-rise concept. Even at 0.3 the composition and character already start shifting toward the example. Around 0.1 my original image finally stays intact, but then the actual concept is barely applied.