[MRG] Fix for float16 overflow on accumulator operations#13010

Merged

rth merged 6 commits intoscikit-learn:masterfrom

baluyotraf:float16-overflow

Jan 26, 2019

Commits on Jan 18, 2019

Fixed overflows on float16 when working with operations involving accumulators (#13007)
baluyotraf
committed

Commits on Jan 19, 2019

Added test for checking StandardScaler float16 overflow (#13007)
baluyotraf
committed
Renamed safe_acc_op to _safe_accumulator_op and moved it to sklearn.utils.extmath. Also fixed some line lengths to fit the 80 limit (#13007)
baluyotraf
committed

Commits on Jan 20, 2019

Commits on Jan 21, 2019

Added a test to verify that both the float64 and float16 has same result with respect to their precisions (#13007)
baluyotraf
committed