nkanazawa1989
commented
8 months ago Sounds like a good first issue :)
Our visualizer uses the bridge pattern to support multiple drawer backends (currently we only have matplotlib though), and the implementation might look complicated. But the curve plotter is a good example to start with. This plotter object must be drawer-backend agnostic and just provides data structure and API for experiments. You need to define a plotter class for every visualization template. You also need to extend the mpl drawer to call box plot API.
boonware
Suggested feature
Perhaps also if Median and Quartiles Including the median and quartiles (Q1 and Q3) to provide insights into the data distribution's central tendency and spread. Box Plots: Generating box plots to visually represent the distribution of the benchmarking results, including the median, quartiles, and potential outliers.