[Request] Dynamic regularization images

Is it possible to have the script to generate the regularization images by some prompts and then then use those to regulate the model being trained?

To generate those images, the system could ignore the model being trained and generate images without using it. I don't know if this is how it works, but maybe in this way the script can even use the same seed that was used for generation for the regulation itself.
With this you could have infinite regularization images, generating a new one whenever "needed" (one for each X amount of steps maybe?).

Thanks in advance!

The issue names no files, tests, or entry points. First identify the training and regularization-image generation paths, then clarify the prompt, seed, and generation-frequency behavior; the work is done when dynamic image generation and its effect on regularization are specified and verified.

Written by the indexing model from the issue text.