How do you run inferences outside of the webui like a script?

[nub2927]

my use case for tensorrt is I need to make a lot of regularization images,
however the way im constructing the regularization images is im drawing from a pool of unique prompts that also correlate to uniqure image sizes necessitating the need for a script that draws from this pool and sets the prompt and images sizes at each inference,
using automatics webui I make use of the custom scripts features to do this but inferences were too slow there,
is there a way for me to make a script that calls the tensort inferences in this same manner?

[ddPn08]

A plugin API will be released soon.