Flatten¶

class Flatten(input_size: Tuple)¶

Bases: bluebird.layers.input.Input

Flatten layer is a type of an Input layer. It transforms Tensor with dimensions (n, m) to a Tensor with dimensions of (n*m, 1).

Example:

input = Flatten((5, 10))
net = NeuralNet([
        input,
        ...
    ])

__init__(input_size: Tuple) → None¶

Initializes the object.

Methods

`__init__`(input_size)	Initializes the object.
`backward`(output)	Reshapes Tensor back to (n, m).
`build`()	Called by the model, before its training step.
`forward`(inputs[, training])	Called each time the data passes throughout the nework.
`get_params`()	Returns params of layer.

backward(output: numpy.ndarray) → numpy.ndarray¶

Reshapes Tensor back to (n, m).

build() → None¶

Called by the model, before its training step.

Prepares the input size for next layer.

forward(inputs: numpy.ndarray, training: bool = False) → numpy.ndarray¶

Called each time the data passes throughout the nework.

Parameters

inputs (Tensor) – input data to the network
training (bool, optional) – set to true during training, and is false when network predicts

Returns

reshaped Tensor (n*m, 1)

Return type

Tensor

get_params() → Dict¶

Returns params of layer.

BlueBird