Module juice::layers::common

Provides common neural network layers.

For now the layers in common should be discribed as layers that are typical layers for building neural networks but are not activation or loss layers.

Re-exports

pub use self::convolution::Convolution;

pub use self::convolution::ConvolutionConfig;

pub use self::dropout::Dropout;

pub use self::dropout::DropoutConfig;

pub use self::linear::Linear;

pub use self::linear::LinearConfig;

pub use self::log_softmax::LogSoftmax;

pub use self::pooling::Pooling;

pub use self::pooling::PoolingConfig;

pub use self::pooling::PoolingMode;

pub use self::rnn::Rnn;

pub use self::rnn::RnnConfig;

pub use self::softmax::Softmax;

Modules

convolution

Convolves the input tensor.

dropout

Applies a Dropout layer to the input data x

linear

Applies a linear transformation to the input data y = a * x + b

log_softmax

Computes the logarithmic softmax of its input.

pooling

Applies pooling to the input.

rnn

Create a Recursive Layer

softmax

Computes the softmax of its input.

Traits

FilterLayer

Provides common utilities for Layers that utilize a filter with stride and padding.