Struct juice::layers::container::sequential::SequentialConfig

pub struct SequentialConfig {
    pub layers: Vec<LayerConfig>,
    pub inputs: Vec<(String, Vec<usize>)>,
    pub force_backward: bool,
}

Specifies configuration parameters for a Sequential Layer.

Fields

layers: Vec<LayerConfig>

Defines the layers of the container via LayerConfigs.

inputs: Vec<(String, Vec<usize>)>

Defines the names and shapes of the input tensors.

The inputs are identified by name so they can be referenced as input tensors in a LayerConfig.

force_backward: bool

Defines if the container will force every layer to do [backpropagation][1]. [1]: https://en.wikipedia.org/wiki/Backpropagation

If set to false, then the execution of backpropagation is determined automatically according to the network structure and learning rates.

Default: false

Implementations

impl SequentialConfig

pub fn find_in_place_output(&self, n: usize) -> Option<String>

Tries to find the output of a previous layer that is usable as in-place output for the n-th layer.

pub fn add_layer(&mut self, layer: LayerConfig)

Add layer at the end of the sequential container.

pub fn add_input(&mut self, input_name: &str, shape: &[usize])

Add a input to the network.

Trait Implementations

impl Clone for SequentialConfig

fn clone(&self) -> SequentialConfig

Returns a copy of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for SequentialConfig

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl Default for SequentialConfig

fn default() -> SequentialConfig

Returns the “default value” for a type.

impl Into<LayerType> for SequentialConfig

fn into(self) -> LayerType

Converts this type into the (usually inferred) input type.