search

byu-dml / metalearn

BYU's python library of useable tools for metalearning
MIT License
22 stars 6 forks source link

Soft timeout #138

Closed bjschoenfeld closed 5 years ago

bjschoenfeld commented 6 years ago

See this issue. A soft timeout means that metafeatures are computed until timeout has been exceeded. Then no more are computed and the computed metafeatures are returned.

The simplest version is to check the total compute time after each metafeature is computed.

A more intricate one would check at the beginning of each call of _get_resource, _get_function, and _get_arguments, etc.

mitar commented 6 years ago

I think you should just compute all metafeatures requested and leave to TA2 to figure out which one to pick to be under a budget.