When queuing generations across multiple models, generate them in order of model.

[ataylorm]

Let's say that I want to generate 10 images with a prompt in 5 different models. It will generate 1, switch model, generate, switch model, etc. This is very inefficient. It would be much better to generate all 10 on a model, then switch the model and generate all queued for that model, etc. This reduces model loading time significantly.