Module view.WeightViewer

Widget that allows the user to view the weights os a neural network in a table format.

Functions

WeightViewer.new () Constructor.
WeightViewer:populate (layer) Populates the model with the connections of a layer.
WeightViewer:prepare (layer_array, bias_array, conn_array) Prepares the widget for new connections.
(private) WeightViewer:build_gui () Builds the interface.
(private) WeightViewer:get_layer_pos (neuron_id) Gets the layer and the position in the layer of a neuron id.
(private) WeightViewer:layer_changed ()
(private) WeightViewer:new_column (label, id)
(private) WeightViewer:row_activated (path) Shows a more detailed description of the connection for the user.
(private) set_search_col (data)


Functions

WeightViewer.new ()
Constructor.

Return value:

New instance of WeightViewer
WeightViewer:populate (layer)
Populates the model with the connections of a layer.

Parameters

  • layer: Layer ("from") to get the connections
WeightViewer:prepare (layer_array, bias_array, conn_array)
Prepares the widget for new connections.

Parameters

  • layer_array: Array with the number of neurons in each layer
  • bias_array: Array with the number of bias in each layer
  • conn_array: Map with the connections
(private) WeightViewer:build_gui ()
Builds the interface.
(private) WeightViewer:get_layer_pos (neuron_id)
Gets the layer and the position in the layer of a neuron id. O(n), where n is the number os layers.

Parameters

  • neuron_id: Neuron id

Return value:

layer, position, is bias
(private) WeightViewer:layer_changed ()
(private) WeightViewer:new_column (label, id)

Parameters

  • label:
  • id:
(private) WeightViewer:row_activated (path)
Shows a more detailed description of the connection for the user. It's implemented this way for performance reasons.

Parameters

  • path:
(private) set_search_col (data)

Parameters

  • data:

Valid XHTML 1.0!