Benchmark¶
How to Run Benchmarks¶
$ cd example
$ python run_benchmark.py -f levy100 -s mcts -m turbo -l threshold_svm -r 40 -t 180
In above example, Levy100 is optimized with LA-MCTS as optimizer. TuRBO sampler and threhold SVM classifier are used. Total 40 runs are executed, each with 180 second time budget, the results will be saved in example/output folder.
The results can then be reported by
$ python benchmark_result.py -f levy100
Only top 10 best runs are used for the report.
