All Versions
42
Latest Version
Avg Release Cycle
70 days
Latest Release
-

Changelog History
Page 1

  • v4.0.0 Changes

    ????-??-??
    • Add Multiple Pole Balancing Environment (#1901, #1951).

    • Added functionality for scaling of data (#1876); see the command-line binding mlpack_preprocess_scale or Python binding preprocess_scale().

    • Add new parameter maximum_depth to decision tree and random forest bindings (#1916).

    • Fix prediction output of softmax regression when test set accuracy is calculated (#1922).

    • Action struct in continuous RL environments now stores the action as a double instead of double[1] (#1941, #1931).

    • Pendulum environment now checks for termination. All RL environments now have an option to terminate after a set number of time steps (no limit by default) (#1941).

    • Add support for probabilistic KDE (kernel density estimation) error bounds when using the Gaussian kernel (#1934).

  • v3.4.2 Changes

    October 28, 2020

    ๐Ÿš€ Released Oct. 28, 2020.

    • โž• Added Mean Absolute Percentage Error.
    • โž• Added Softmin activation function as layer in ann/layer.
    • Fix spurious ARMA_64BIT_WORD compilation warnings on 32-bit systems (#2665).
  • v3.4.1 Changes

    September 07, 2020

    ๐Ÿš€ Released Sep. 7, 2020.

    ๐Ÿ›  Fix incorrect parsing of required matrix/model parameters for command-line bindings (#2600).

    โž• Add manual type specification support to data::Load() and data::Save() (#2084, #2135, #2602).

    โœ‚ Remove use of internal Armadillo functionality (#2596, #2601, #2602).

  • v3.4.0 Changes

    September 01, 2020

    ๐Ÿš€ Released Sept. 1st, 2020.

    โš  Issue warnings when metrics produce NaNs in KFoldCV (#2595).

    Added bindings for R during Google Summer of Code (#2556).

    โž• Added common striptype function for all bindings (#2556).

    ๐Ÿ”จ Refactored common utility function of bindings to bindings/util (#2556).

    Renamed InformationGain to HoeffdingInformationGain in methods/hoeffding_trees/information_gain.hpp (#2556).

    โž• Added macro for changing stream of printing and warnings/errors (#2556).

    โž• Added Spatial Dropout layer (#2564).

    ๐Ÿ‘ฎ Force CMake to show error when it didn't find Python/modules (#2568).

    ๐Ÿ”จ Refactor ProgramInfo() to separate out all the different information (#2558).

    โž• Add bindings for one-hot encoding (#2325).

    โž• Added Soft Actor-Critic to RL methods (#2487).

    โž• Added Categorical DQN to q_networks (#2454).

    โž• Added N-step DQN to q_networks (#2461).

    โž• Add Silhoutte Score metric and Pairwise Distances (#2406).

    โž• Add Go bindings for some missed models (#2460).

    Replace boost program_options dependency with CLI11 (#2459).

    โž• Additional functionality for the ARFF loader (#2486); use case sensitive categories (#2516).

    Add bayesian_linear_regression binding for the command-line, Python, Julia, and Go. Also called "Bayesian Ridge", this is equivalent to a version of linear regression where the regularization parameter is automatically tuned (#2030).

    ๐Ÿ›  Fix defeatist search for spill tree traversals (#2566, #1269).

    ๐Ÿ›  Fix incremental training of logistic regression models (#2560).

    Change default configuration of BUILD_PYTHON_BINDINGS to OFF (#2575).

  • v3.3.2

    June 18, 2020
  • v3.3.1 Changes

    April 30, 2020

    ๐Ÿš€ Released April 29th, 2020.

    ๐Ÿ“š Minor Julia and Python documentation fixes (#2373).

    โšก๏ธ Updated terminal state and fixed bugs for Pendulum environment (#2354, #2369).

    โž• Added EliSH activation function (#2323).

    โž• Add L1 Loss function (#2203).

    Pass CMAKE_CXX_FLAGS (compilation options) correctly to Python build (#2367).

    ๐Ÿ”ฆ Expose ensmallen Callbacks for sparseautoencoder (#2198).

    ๐Ÿ›  Bugfix for LARS class causing invalid read (#2374).

    โž• Add serialization support from Julia; use mlpack.serialize() and mlpack.deserialize() to save and load from IOBuffers.

  • v3.3.0 Changes

    April 07, 2020

    ๐Ÿš€ Released April 7th, 2020.

    Templated return type of Forward function of loss functions (#2339).

    โž• Added R2 Score regression metric (#2323).

    โž• Added mean squared logarithmic error loss function for neural networks (#2210).

    โž• Added mean bias loss function for neural networks (#2210).

    ๐Ÿ—„ The DecisionStump class has been marked deprecated; use the DecisionTree class with NoRecursion=true or use ID3DecisionStump instead (#2099).

    โž• Added probabilities_file parameter to get the probabilities matrix of AdaBoost classifier (#2050).

    ๐Ÿ›  Fix STB header search paths (#2104).

    โž• Add DISABLE_DOWNLOADS CMake configuration option (#2104).

    โž• Add padding layer in TransposedConvolutionLayer (#2082).

    ๐Ÿ›  Fix pkgconfig generation on non-Linux systems (#2101).

    ๐ŸŒฒ Use log-space to represent HMM initial state and transition probabilities (#2081).

    โž• Add functions to access parameters of Convolution and AtrousConvolution layers (#1985).

    โž• Add Compute Error function in lars regression and changing Train function to return computed error (#2139).

    Add Julia bindings (#1949). Build settings can be controlled with the BUILD_JULIA_BINDINGS=(ON/OFF) and JULIA_EXECUTABLE=/path/to/julia CMake parameters.

    CMake fix for finding STB include directory (#2145).

    Add bindings for loading and saving images (#2019); mlpack_image_converter from the command-line, mlpack.image_converter() from Python.

    โž• Add normalization support for CF binding (#2136).

    โž• Add Mish activation function (#2158).

    โšก๏ธ Update init_rules in AMF to allow users to merge two initialization rules (#2151).

    โž• Add GELU activation function (#2183).

    ๐Ÿ‘ Better error handling of eigendecompositions and Cholesky decompositions (#2088, #1840).

    โž• Add LiSHT activation function (#2182).

    โž• Add Valid and Same Padding for Transposed Convolution layer (#2163).

    โž• Add CELU activation function (#2191)

    โž• Add Log-Hyperbolic-Cosine Loss function (#2207)

    ๐Ÿ”„ Change neural network types to avoid unnecessary use of rvalue references (#2259).

    โฌ†๏ธ Bump minimum Boost version to 1.58 (#2305).

    ๐Ÿ”จ Refactor STB support so HAS_STB macro is not needed when compiling against mlpack (#2312).

    โž• Add Hard Shrink Activation Function (#2186).

    โž• Add Soft Shrink Activation Function (#2174).

    โž• Add Hinge Embedding Loss Function (#2229).

    โž• Add Cosine Embedding Loss Function (#2209).

    โž• Add Margin Ranking Loss Function (#2264).

    ๐Ÿ›  Bugfix for incorrect parameter vector sizes in logistic regression and softmax regression (#2359).

  • v3.2.2 Changes

    November 26, 2019

    ๐Ÿš€ Released Nov. 26th, 2019.

    • โž• Add valid and same padding option in Convolution and Atrous Convolution layer (#1988).
    • โž• Add Model() to the FFN class to access individual layers (#2043).
    • ๐Ÿ“š Update documentation for pip and conda installation packages (#2044).
    • Add bindings for linear SVM (#1935); mlpack_linear_svm from the command-line, linear_svm() from Python.
    • โž• Add support to return the layer name as std::string (#1987).
    • Speed and memory improvements for the Transposed Convolution layer (#1493).
    • ๐Ÿ›  Fix Windows Python build configuration (#1885).
    • Validate md5 of STB library after download (#2087).
    • ๐Ÿ”– Add __version__ to __init__.py (#2092).
  • v3.2.1

    October 01, 2019
  • v3.2.0 Changes

    September 26, 2019

    ๐Ÿš€ Released Sept. 25, 2019.

    ๐Ÿ›  Fix occasionally-failing RADICAL test (#1924).

    ๐Ÿ›  Fix gcc 9 OpenMP compilation issue (#1970).

    โž• Added support for loading and saving of images (#1903).

    โž• Add Multiple Pole Balancing Environment (#1901, #1951).

    Added functionality for scaling of data (#1876); see the command-line binding mlpack_preprocess_scale or Python binding preprocess_scale().

    โž• Add new parameter maximum_depth to decision tree and random forest bindings (#1916).

    ๐Ÿ›  Fix prediction output of softmax regression when test set accuracy is calculated (#1922).

    0๏ธโƒฃ Pendulum environment now checks for termination. All RL environments now have an option to terminate after a set number of time steps (no limit by default) (#1941).

    โž• Add support for probabilistic KDE (kernel density estimation) error bounds when using the Gaussian kernel (#1934).

    ๐Ÿ›  Fix negative distances for cover tree computation (#1979).

    ๐Ÿ›  Fix cover tree building when all pairwise distances are 0 (#1986).

    ๐Ÿ‘Œ Improve KDE pruning by reclaiming not used error tolerance (#1954, #1984).

    ๐Ÿ“œ Optimizations for sparse matrix accesses in z-score normalization for CF (#1989).

    Add kmeans_max_iterations option to GMM training binding gmm_train_main.

    โฌ†๏ธ Bump minimum Armadillo version to 8.400.0 due to ensmallen dependency requirement (#2015).