I encountered strange behavior with the C bindings when I attempted to use LSH (I got random memory written out to my match indices). It seems that Harris distance + LSH doesn't work with the C bindings. Would there be any issue with me adding Harris distance appropriately to the C bindings. Also, was the behavior I encountered unexpected and was it supposed to work?
For clarification, I called flann_find_nearest_neighbors_byte with single byte input, 1 row, 1 column. It wrote back these: 1836216064, 1 (when it should have been 0), 52430598, 50462976, 0 (when it should have been 1), 18446744071676166400, and 1718026496. It seems to be giving me garbage memory, since some of those numbers look like stack garbage.
I encountered strange behavior with the C bindings when I attempted to use LSH (I got random memory written out to my match indices). It seems that Harris distance + LSH doesn't work with the C bindings. Would there be any issue with me adding Harris distance appropriately to the C bindings. Also, was the behavior I encountered unexpected and was it supposed to work?
For clarification, I called
flann_find_nearest_neighbors_byte
with single byte input, 1 row, 1 column. It wrote back these:1836216064
,1
(when it should have been 0),52430598
,50462976
,0
(when it should have been1
),18446744071676166400
, and1718026496
. It seems to be giving me garbage memory, since some of those numbers look like stack garbage.