search

mlpack / benchmarks

Machine Learning Benchmark Scripts
101 stars 49 forks source link

Classification Algorithms in R #104

Closed Iron-Stark closed 7 years ago

zoq commented 7 years ago

If you like, I can resolve the merge conflict for you just let me know.

Iron-Stark commented 7 years ago

@mlpack-jenkins test this

zoq commented 7 years ago

Can you give the DecisionTree_R_TEST(unittest.TestCase): test more time and try again.

Iron-Stark commented 7 years ago

@zoq

What happens here is certain packages require dependencies installed which can take time to download and setup. Should I increase the timeouts for all the tests because depending on the net speed of the user running these packages on a local system the tests might fail if the net connection is slow. Please let me know what you think about this.

zoq commented 7 years ago

Currently it looks like everytime we run a benchmark script, we reinstall the packages, is there a way to check for the package first? I think for the R script it's reasonable to slightly increase the timeout, but I'm not sure we should increase the timeout for all tests.

rcurtin commented 7 years ago

Another approach which might be better is to install all necessary R packages in install_R.sh; then there would be no issue with test timeouts.

zoq commented 7 years ago

Agreed, that is a better solution to solve the timeout issues.

rcurtin commented 7 years ago

@mlpack-jenkins test this

zoq commented 7 years ago

@mlpack-jenkins test this

zoq commented 7 years ago

As @rcurtin already said, nice work!