[Feature Request]: use previously cached IR models

Is there an existing issue for this?

[X] I have searched the existing issues and checked the recent builds/commits

What would your feature do ?

I am currently testing this repo with a single model and found that after restarting the app(shutting down and restarting again from CLI, did not try from webui), the script creates the model again even if I load previous settings(including the prompt itself), so the suggestion I have is to store multiple preconfigured IR models.

Proposed workflow

adding a new tab in the webUI that will create configurations.
in that tab, we set resolution, batch size, scheduler, device(cpu/gpu) etc.
there will be a button that will create the required IR file according to the configuration.
in the text2img tab, when we select accelerate with OpenVINO, it will have an option to use premade IR model.
selecting the premade IR model will automatically sets batch size, resolution, etc. (or maybe also disables them)
if someone even restarts the computer, it will have the option to load a premade IR model.

Additional information

adding a new tab to create configurations is not important you can just add a checkbox "save this IR model" that will enable a textfield to name the configuration or something similar.

openvinotoolkit / stable-diffusion-webui