ConvLSTM3D

Description

Adds the weights of the ConvLSTM3D layer to the weights table. Type : polymorphic.

Input parameters

 index : integer, index of layer.
 kernel : array, 5D values. kernel = [4*n_filters, channels, size[0], size[1], size[2]].
 recurrent_kernel : array, 5D values.  recurrent_kernel = [4*n_filters, n_filters, size[0], size[1], size[2]].
 bias : array, 1D values. bias = [4*n_filters].

Output parameters

Dimension

• kernel = [4*n_filters, channels, size[0], size[1], size[2]]

The kernel size depends on the input of the ConvLSTM3D layer and the parameters n_filters and size of the ConvLSTM3D cell.
For example, if the input of the layer has a size of [samples = 10, time = 8, channels =  5, rows = 4, cols = 3, depth = 2], n_filters a value of 6 and size the value [3, 3, 3], then kernel will have a size of [4*n_filters = 6, channels = 5, size[0] = 3, size[1] = 3, size[2] = 3].

• recurrent_kernel = [4*n_filters, n_filters, size[0], size[1], size[2]]

The size of recurrent_kernel depends on the parameters n_filters and size of the ConvLSTM3D cell.
For example, if n_filters has a value of 6 and size the value [3, 3, 3], then recurrent_kernel will have a size of [4*n_filters = 6, n_filters = 6, size[0] = 3, size[1] = 3, size[2] = 3].

• bias = [4*n_filters]

The size of bias depends on the parameter n_filters of the ConvLSTM3D cell.
For example if n_filters has a value of 6 then the bias size will be [4*n_filters = 6].

Example

All these exemples are snippets PNG, you can drop these Snippet onto the block diagram and get the depicted code added to your VI (Do not forget to install Deep Learning library to run it).