high up is a concept that Anima has already learned. For this LoRA, I used 100 training images that clearly convey the subject being in a "high place" within the composition. Examples include rooftops, cityscapes with buildings, above the clouds, and floating.Conversely, images where the character was too close-up to clearly tell that they were high above the ground were excluded from the training set.Applying this LoRA may improve the sense of height and elevation, but in some cases, the result may change very little.
