nn_utils_weight_norm: nn_utils_weight_norm

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

nn_utils_weight_norm

Description

Applies weight normalization to a parameter in the given module.

Details

    \eqn{\mathbf{w} = g \dfrac{\mathbf{v}}{\|\mathbf{v}\|}}

Weight normalization is a reparameterization that decouples the magnitude of a weight tensor from its direction. This replaces the parameter specified by name (e.g. 'weight') with two parameters: one specifying the magnitude (e.g. 'weight_g') and one specifying the direction (e.g. 'weight_v').

Value

The original module with the weight_v and weight_g paramters.

Methods

Public methods

• nn_utils_weight_norm$new()

• nn_utils_weight_norm$compute_weight()

• nn_utils_weight_norm$apply()

• nn_utils_weight_norm$call()

• nn_utils_weight_norm$recompute()

• nn_utils_weight_norm$remove()

• nn_utils_weight_norm$clone()

---

Method new()

Usage

nn_utils_weight_norm$new(name, dim)

Arguments

name

(str, optional): name of weight parameter

dim

(int, optional): dimension over which to compute the norm

---

Method compute_weight()

Usage

nn_utils_weight_norm$compute_weight(module, name = NULL, dim = NULL)

Arguments

module

(Module): containing module

name

(str, optional): name of weight parameter

dim

(int, optional): dimension over which to compute the norm

---

Method apply()

Usage

nn_utils_weight_norm$apply(module, name = NULL, dim = NULL)

Arguments

module

(Module): containing module

name

(str, optional): name of weight parameter

dim

(int, optional): dimension over which to compute the norm

---

Method call()

Usage

nn_utils_weight_norm$call(module)

Arguments

module

(Module): containing module

---

Method recompute()

Usage

nn_utils_weight_norm$recompute(module)

Arguments

module

(Module): containing module

---

Method remove()

Usage

nn_utils_weight_norm$remove(module, name = NULL)

Arguments

module

(Module): containing module

name

(str, optional): name of weight parameter

---

Method clone()

The objects of this class are cloneable with this method.

Usage

nn_utils_weight_norm$clone(deep = FALSE)

Arguments

deep

Whether to make a deep clone.

Note

The pytorch Weight normalization is implemented via a hook that recomputes the weight tensor from the magnitude and direction before every forward() call. Since torch for R still do not support hooks, the weight recomputation need to be done explicitly inside the forward() definition trough a call of the recompute() method. See examples.

By default, with dim = 0, the norm is computed independently per output channel/plane. To compute a norm over the entire weight tensor, use dim = NULL.

@references https://arxiv.org/abs/1602.07868

Examples

if (torch_is_installed()) {
x = nn_linear(in_features = 20, out_features = 40)
weight_norm = nn_utils_weight_norm$new(name = 'weight', dim = 2)
weight_norm$apply(x)
x$weight_g$size()
x$weight_v$size()
x$weight

# the recompute() method recomputes the weight using g and v. It must be called
# explicitly inside `forward()`.
weight_norm$recompute(x)

}

torch documentation built on May 29, 2024, 9:54 a.m.