mirror of https://github.com/davisking/dlib.git
Made spec more clear
--HG-- extra : convert_revision : svn%3Afdd8eb12-d10e-0410-9acb-85c331704f74/trunk%403576
This commit is contained in:
parent
f505d4d954
commit
93971a811a
|
@ -41,7 +41,7 @@ namespace dlib
|
|||
|
||||
|
||||
At the end of the for loop you will have your random subset of 1000 samples. And by
|
||||
random I mean that each of the 1000000 data samples has an equal change of ending
|
||||
random I mean that each of the 1000000 data samples has an equal chance of ending
|
||||
up in the rand_subset object.
|
||||
|
||||
|
||||
|
@ -61,7 +61,12 @@ namespace dlib
|
|||
In the above example we only actually fetch the data sample into memory if we
|
||||
know that the rand_subset would include it into the random subset. Otherwise,
|
||||
we can just call the empty add().
|
||||
|
||||
|
||||
Finally, note that the random_subset_selector uses a deterministic pseudo-random
|
||||
number generator under the hood. Moreover, the default constructor always seeds
|
||||
the random number generator in the same way. So unless you call set_seed()
|
||||
each instance of the random_subset_selector will function identically.
|
||||
!*/
|
||||
public:
|
||||
typedef T type;
|
||||
|
|
Loading…
Reference in New Issue