Issue warnings in multistart_hyperparameter_optimization of python version

Yelp / MOE

A global, black box optimization engine for real world metric optimization.

Other

1.3k stars 139 forks source link

Issue warnings in multistart_hyperparameter_optimization of python version #441

Open yf275 opened 9 years ago

yf275 commented 9 years ago

Hi,

For the one in optimal_learning/python/python_version/log_likelihood.py, Line 107 is trying to apply numpy.log10 on hyperparameter_optimizer.domain._domain_bounds. I think we should issue a warning to users when they include negative values in domain_bounds so numpy.log10 would return a number instead of nan Similar thing for Line 128 of log_likelihood.py in optimal_learning/python/cpp_wrappers/log_likelihood.py.

suntzu86 commented 8 years ago

Sorry I've been gone for so long! See: #448 for details.

This is similar to https://github.com/Yelp/MOE/issues/446, I think. Sorry, yes MOE's input validity checking isn't very thorough.