search

mlpack / benchmarks

Machine Learning Benchmark Scripts
101 stars 49 forks source link

The benchmark of MLPack/decision_tree has been done #115

Closed rickllyxu closed 6 years ago

rickllyxu commented 6 years ago

I runned the command make run BLOCK=mlpack METHODBLOCK=DTC. It is successfully executed.

mlpack-jenkins commented 6 years ago

Can one of the admins verify this patch?

rcurtin commented 6 years ago

@mlpack-jenkins test this please

rickllyxu commented 6 years ago

I am not sure I submited changes successfully. I revisioned serveral code and done some test, such as make run BLOCK=mlpack METHODBLOCK=DTC and make test. Here is the output of former command: [INFO ] CPU Model: Intel(R) Core(TM) i5-3210M CPU @ 2.50GHz [INFO ] Distribution: debian stretch/sid [INFO ] Platform: x86_64 [INFO ] Memory: 7.6943359375 GB [INFO ] CPU Cores: 4

dataset mlpack
iris 0.000399
oilspill 0.027024
scene 2.074252
webpage 14.852665
isolet 21.783903
mammography 0.752401
reuters 40.933589
abalone19 0.031373
sickEuthyroid 0.024043
abalone7 -2
satellite 0.082406
ecoli 0.002451

Because there is not a file named abalone7_test.csv under datasets, so benchmarking on abalone7 failed. And I cannot execute scikit/dtc.py benchmarking because compiler cannot find timeout_decorator, though I have found zoq added timeout_decorator in commit log.

UPDATE Here I add the result of benchmarking on abalone7. Seems to there is a gap bettween the following results and above.

dataset mlpack
iris 0.000392
oilspill 0.027961
scene 2.032154
webpage 13.778425
isolet 21.953052
mammography 0.734820
reuters 41.359781
abalone19 0.032652
sickEuthyroid 0.025288
abalone7 0.052745
satellite 0.081667
ecoli 0.000303
rcurtin commented 6 years ago

Ok, I will go ahead and merge this now then. Thanks again!

If you want, I can mail you some mlpack stickers that you can put on your laptop. If you'd like some, just send me an email at ryan@ratml.org with your mailing address and I will put them in the mail.