random_channel

Random Channel Layer API

This module provides classes to build random channel layers.

Classes

Name	Description
RandomChannel	Random channel

Machine-readable model

• RandomChannelclassRandomly picks a single channel from the input samples.

class

RandomChannel

Python

Randomly picks a single channel from the input samples.

helia_edge/layers/preprocessing/random_channel.py:17

RandomChannel(batchwise: bool = False, **kwargs={})

Randomly picks a single channel from the input samples.

Base class: BaseAugmentation.

Inherited from BaseAugmentation: augment_masks(), augment_targets(), batch_augment(), call().

Parameters

Parameters of RandomChannel

Name	Type	Default	Description
batchwise	bool	False	If True, grabs same channel from all samples in batch.

attribute

batchwise

Python

helia_edge/layers/preprocessing/random_channel.py:29

batchwise: bool = batchwise

method

get_random_transformations

Python

Generate random channel index.

helia_edge/layers/preprocessing/random_channel.py:31

get_random_transformations(input_shape)

Generate random channel index.

method

augment_samples

Python

Augment samples

helia_edge/layers/preprocessing/random_channel.py:50

augment_samples(inputs)

Augment samples

method

augment_sample

Python

Augment a sample during training.

helia_edge/layers/preprocessing/random_channel.py:62

augment_sample(inputs) -> keras.KerasTensor

Augment a sample during training.

method

get_config

Python

helia_edge/layers/preprocessing/random_channel.py:68

get_config()