AdaGrad

Constructors

this

this(ulong num_epochs, float lr, ulong mini_batch_sz, float eps_)

Undocumented in source.

Members

Functions

current_lr

float current_lr(size_t k, size_t j)

Undocumented in source. Be warned that the author may not have intended to support it.

dup

AdaGrad dup()

Undocumented in source. Be warned that the author may not have intended to support it.

params_to_optimize

void params_to_optimize(float[][] W_, float[][] grad_)

Undocumented in source. Be warned that the author may not have intended to support it.

register

void register(NeuralLayer layer)

Undocumented in source. Be warned that the author may not have intended to support it.

toString

string toString()

Undocumented in source. Be warned that the author may not have intended to support it.

update

void update(NeuralLayer layer, float[] ext_grad)

Undocumented in source. Be warned that the author may not have intended to support it.

update

void update(NeuralLayer layer, SparseF[] ext_grad)

Undocumented in source. Be warned that the author may not have intended to support it.

update_general

void update_general(NeuralLayer layer, V[] ext_grad)

Undocumented in source. Be warned that the author may not have intended to support it.

Mixins

__anonymous

mixin opCallNew

Undocumented in source.

Static functions

ada_op1

void ada_op1(float[] row, float[] g)

Undocumented in source. Be warned that the author may not have intended to support it.

ada_op2

void ada_op2(float[] row, float[] g, float[] sg, float lr_, float eps_)

Undocumented in source. Be warned that the author may not have intended to support it.

update_matrix

void update_matrix(float[][] w, float[][] g, float[][] sum_g_sq, float lr_, float eps_)

Undocumented in source. Be warned that the author may not have intended to support it.

Variables

W

float[][] W;

Undocumented in source.

_update

void delegate(NeuralLayer, float[]) _update;

Undocumented in source.

acc_grad

float[][] acc_grad;

Undocumented in source.

eps

float eps;

Undocumented in source.

grad

float[][] grad;

Undocumented in source.

Mixed In Members

From mixin opCallNew

opCall

auto opCall(T args)

Undocumented in source. Be warned that the author may not have intended to support it.

Inherited Members

From SGDOptimizer

num_epochs

ulong num_epochs;

Undocumented in source.

mini_batch_sz

ulong mini_batch_sz;

Undocumented in source.

lr

float lr;

Undocumented in source.

cnt

ulong cnt;

Undocumented in source.

priors

AdditiveLinearPrior[] priors;

Undocumented in source.

prox

ProxyLinearPrior prox;

Undocumented in source.

learn

void learn(NeuralNet nn, D data, S delegate(R net_out, ref T ex, ref V[] grad) grad_f, bool verbose, uint num_cores)

Undocumented in source. Be warned that the author may not have intended to support it.

current_lr

float current_lr(size_t k, size_t j)

Undocumented in source.

register

void register(NeuralLayer layer)

Undocumented in source.

update

void update(NeuralLayer layer, float[] ext_grad)

Undocumented in source.

update

void update(NeuralLayer layer, SparseF[] ext_grad)

Undocumented in source.

dup

Optimizer dup()

Undocumented in source.

toString

string toString()

Undocumented in source. Be warned that the author may not have intended to support it.

auto nn = NeuralNet()
    .stack(SparseData(1000))
    .stack(Linear(1));

// Adagrad learning with 5 epochs, learning rate 0.1, mini-batch size of 100:
nn.learn(data, "square", AdaGrad(5, 0.1, 100));

• ADAM
• AdaGrad
• SGDOptimizer
• ShadowSGDOptimizer

• Optimizer