For small files or the individual rule timings, the current millisecond precision (which does not display any decimals) can be a bit too coarse, if results are all just 0 or 1.
Microsecond precision would be better. The information is already there, captured in nanoseconds, so it's just about the display in the UI.
I provided a patch over in my fork alexkli@6c7199e which results in this:

Note the Time column as well as Parse time and Prediction time fields on the right.
You can try out with the 1.8.4-alexkli build from my fork.
(Note that providing a pull request is a bit tricky since I had to make some changes to the project setup on my branch before it worked for me, and I made another change for #322 on the same branch. I am happy to rebase things for a PR if accepted.)