I have created a flow with a custom model for extracting 7 fields from a standard document. The issue is that the extract information from documents action takes 17 seconds or more to complete. This is too long for my users, I'm concerned that they will think the system froze and will refresh/exit the application.
The custom model has an accuracy score of 99%. I have tried adding more documents for training but this does not speed up the model. I have looked online but cannot find any documentation to show me how to speed up the model. What can I do to speed it up? My goal is to have it run under 2 seconds.
Here is a screen shot of the flow.