I think it could be useful to get a thread going to share useful methods and techniques we've found while generating images.

for instance I've found tiled diffusion extremely useful for increasing the resolution and improving the details of images afteri select some good ones from a batch: https://github.com/pkuliyi2015/multidiffusion-upscaler-for-automatic1111

I've also been getting a ton of use out of the jinja2 templates feature of dynamic prompts https://github.com/adieyal/sd-dynamic-prompts/blob/main/jinja2.md I've been using it to test a wide variety of different prompts, situations, and characters in a large batch overnight.

recently ive been trying to use depthmaps with controlnet to create better posed images, especially useful for stuff like impalement. https://github.com/Mikubill/sd-webui-controlnet I've been generating the depthmaps from guro images I have saved(3d renders work well), and also doing poses using https://posemy.art/ mixed luck so far.