nn_prune_head: Prune top layer(s) of a network

In torch: Tensors and Neural Networks with 'GPU' Acceleration

Prune top layer(s) of a network

Description

Prune head_size last layers of a nn_module in order to replace them by your own head, or in order to use the pruned module as a sequential embedding module.

Usage

nn_prune_head(x, head_size)

Arguments

x	nn_network to prune
head_size	number of nn_layers to prune

Value

a nn_sequential network with the top nn_layer removed

Examples

if (torch_is_installed()) {
if (torch_is_installed()) {
x <- nn_sequential(
  nn_relu(),
  nn_tanh(),
  nn_relu6(),
  nn_relu(),
  nn_linear(2,10),
  nn_batch_norm1d(10),
  nn_tanh(),
  nn_linear(10,3)
)  
prune <- nn_prune_head(x, 3)
prune
}
}

torch documentation built on June 7, 2023, 6:19 p.m.