[mlpack-git] [mlpack] Softmax enhance (#455)

stereomatchingkiss notifications at github.com
Sun Sep 27 07:20:08 EDT 2015

Enhance the class SoftMaxRegression

1 : fix bugs--constructor did not initialize fitIntercept(the one which accept optimizer)
2 : add default constructor
3 : add function train
4 : add function Serialize
You can view, comment on, or merge this pull request online at:


-- Commit Summary --

  * improve speed of SparseAutoencoder and make it more flexible
  * fix example of sparse_autoencoder
  * reduce the cost of construction and destruction of the layers
  * refine the example; the size of the basis should not too big
  * refine the example
  * refine the example
  * can use Greedy parameters to determine the Gradient should omit some repeated calculation or not
  * add test cases of modified SparseAutoEncoder
  * add test
  * change the name of test suite to avoid definition confliction
  * change boost test header to avoid build issue
  * prevent nan/inf values
  * first commit
  * 1 : fix bugs--constructor did not initialize fitIntercept

-- File Changes --

    A src/export_filters.cpp (104)
    A src/export_filters.hpp (5)
    A src/mlpack/methods/ann/activation_functions/lazy_logistic_function.hpp (90)
    A src/mlpack/methods/ann/sparse_autoencoder_function.hpp (409)
    M src/mlpack/methods/softmax_regression/softmax_regression.hpp (97)
    M src/mlpack/methods/softmax_regression/softmax_regression_impl.hpp (92)
    M src/mlpack/tests/CMakeLists.txt (1)
    A src/mlpack/tests/sparse_autoencoder_test_2.cpp (278)

-- Patch Links --


Reply to this email directly or view it on GitHub:
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailman.cc.gatech.edu/pipermail/mlpack-git/attachments/20150927/a387d425/attachment.html>

More information about the mlpack-git mailing list