additive iterative image generation
It would be a useful feature have an option to be able to start with simple prompts and add to them, and build on top of the same image, so if we find parts we like, we can keep them, instead of generating an entirely new image each time. For example, the following images had good results for each individual element added to the prompt, but each time completely changed the composition. The final image is the key elements arranged how I originally wanted them to appear. 



