quickdudley/eocalon - Change TCRWQZ2CDTFKYJOYC5G46BYTADRPXXBVZMCTXF2PJRKURLLZ4VGQC

Model trainer: use less preferred activation function because tfjs can deserialize it

Created by quickdudley on July 7, 2021

TCRWQZ2CDTFKYJOYC5G46BYTADRPXXBVZMCTXF2PJRKURLLZ4VGQC

Dependencies

In channels

main

generic

Change contents

Replacement in trainmodel/src/model.py at line 7 [4.1123]

B:BD[3.26] → [3.26:376]

    layer1 = keras.layers.Dense(20, activation=tf.math.asinh)(inputs)
    layer2 = keras.layers.Dense(20, activation=tf.math.asinh)(layer1)
    layer3 = keras.layers.Dense(20, activation=tf.math.asinh)(keras.layers.concatenate([layer1,layer2]))
    layer4 = keras.layers.Dense(20, activation=tf.math.asinh)(keras.layers.concatenate([layer1,layer2]))

[3.26]

[2.0]

    layer1 = keras.layers.Dense(20, activation=keras.activations.softplus)(inputs)
    layer2 = keras.layers.Dense(20, activation=keras.activations.softplus)(layer1)
    layer3 = keras.layers.Dense(20, activation=keras.activations.softplus)(keras.layers.concatenate([layer1,layer2]))
    layer4 = keras.layers.Dense(20, activation=keras.activations.softplus)(keras.layers.concatenate([layer1,layer2]))

Replacement in trainmodel/src/model.py at line 14 [4.1123]

∅:D[2.349] → [3.725:1312]

B:BD[3.725] → [3.725:1312]

    layer8 = keras.layers.Dense(20, activation=tf.math.asinh)(keras.layers.concatenate([layer3,layer5,layer7]))
    layer9 = keras.layers.Dense(20, activation=tf.math.asinh)(keras.layers.concatenate([layer4,layer6,layer7]))
    layer10 = keras.layers.Dense(20, activation=tf.math.asinh)(keras.layers.concatenate([layer7,layer8,layer9]))
    layer11 = keras.layers.Dense(20, activation=tf.math.asinh)(keras.layers.concatenate([layer1,layer3,layer8,layer10]))
    layer12 = keras.layers.Dense(20, activation=tf.math.asinh)(keras.layers.concatenate([layer1,layer4,layer9,layer10,layer11]))

[2.349]

[3.1312]

    layer8 = keras.layers.Dense(20, activation=keras.activations.softplus)(keras.layers.concatenate([layer3,layer5,layer7]))
    layer9 = keras.layers.Dense(20, activation=keras.activations.softplus)(keras.layers.concatenate([layer4,layer6,layer7]))
    layer10 = keras.layers.Dense(20, activation=keras.activations.softplus)(keras.layers.concatenate([layer7,layer8,layer9]))
    layer11 = keras.layers.Dense(20, activation=keras.activations.softplus)(keras.layers.concatenate([layer1,layer3,layer8,layer10]))
    layer12 = keras.layers.Dense(20, activation=keras.activations.softplus)(keras.layers.concatenate([layer1,layer4,layer9,layer10,layer11]))