search

lf-lang / benchmarks-lingua-franca

2 stars 4 forks source link

Modified runner script for continuous benchmarking #2

Closed lhstrh closed 2 years ago

cmnrd commented 2 years ago

Can you elaborate on why this is needed and what it is doing? Why is the csv not sufficient?

cmnrd commented 2 years ago

After checking the changes again, I realized that apparently the json export circumvents hydra's directory structure. I don't think that is a good idea, as the necessary assumptions are easily violated. Note that there is the collect_results.py script for merging the csv outputs of all the individual benchmark runs. If you really need the json output and cannot use the csv, this script would be a better place to make the change.