LoRA training example gen settings that fit the base model

The settings used for example gens when training LoRASs are only really good for SD1.5. I know example gens are just to get a good idea of the model’s output… But if I want to get a good idea of what a model’s output will be (such as a Flux LoRA), a 512×512 image with sd1.5 sampler/scheduler/cfg doesn’t help.