Denser

Grammar: Genetics_ANN/src/cnn.grammar.txt at 8b4206d60a79e8da51f726a309fa1f4c924a4a90. phase: 0=image features, 1=flattened classification, 2=ten-class head, 3=learning configuration. padding: 0=valid, 1=same; activation: 0=tanh, 1=relu, 2=sigmoid; pooling: 0=average, 1=max.
Table of Contents

Types

Denser

(inductive)
Grammar: Genetics_ANN/src/cnn.grammar.txt at 8b4206d60a79e8da51f726a309fa1f4c924a4a90. phase: 0=image features, 1=flattened classification, 2=ten-class head, 3=learning configuration. padding: 0=valid, 1=same; activation: 0=tanh, 1=relu, 2=sigmoid; pooling: 0=average, 1=max.
inductive Denser
Measures
phase : Int
dataset : Int
side : Int
channels : Int
features : Int
hidden : Int

Constructors

For each type in this module: inductive types list their declared constructors; opaque types list any local function whose name starts with mk and which returns a value of that type.

Denser (inductive)

input

constructor
{ n : Denser | phase n = 0 && dataset n = d && features n = 0 && hidden n = 0 && (d = 0 && side n = 28 && channels n = 1 || d = 1 && side n = 32 && channels n = 3) }

conv

constructor
{ r : Denser | phase r = 0 && dataset r = dataset n && side r = out && channels r = filters && features r = features n + 1 && hidden r = 0 }

pool

constructor
{ r : Denser | phase r = 0 && dataset r = dataset n && side r = out && channels r = channels n && features r = features n + 1 && hidden r = 0 }

flatten

constructor
{r : Denser | phase r = 1 && dataset r = dataset n && features r = features n && hidden r = 0}

dense

constructor
{r : Denser | phase r = 1 && dataset r = dataset n && features r = features n && hidden r = hidden n + 1}

softmax

constructor
{r : Denser | phase r = 2 && dataset r = dataset n}

sgd

constructor
{r : Denser | phase r = 3 && dataset r = dataset n}

adam

constructor
{r : Denser | phase r = 3 && dataset r = dataset n}

rmsprop

constructor
{r : Denser | phase r = 3 && dataset r = dataset n}