[feature request] add buttons to load and unload models from vram

this extension is pretty cool but it is very annoying to use in a1111 because it automatically loads models into vram on startup and the only way to unload them is to completely disable the extension.

[aihao2000]

@Hugs288 Thank you for your attention. I have added the function of switch device in the new version. By default, it use cpu and does not occupy vram. Click to switch to cuda or switch to cpu.