Custom Function

Applies a user-supplied function to a tensor.

Parameters

By default, these are inferred as all but the first arguments of the function fn. It is also possible to specify these more explicitly via the param_set constructor argument.

Input and Output Channels

One input channel called "input" and one output channel called "output". For an explanation see PipeOpTorch.

Super classes

mlr3pipelines::PipeOp -> mlr3torch::PipeOpTorch -> PipeOpTorchFn

Methods

Public methods

• PipeOpTorchFn$new()

• PipeOpTorchFn$clone()

Inherited methods

• mlr3pipelines::PipeOp$help()
• mlr3pipelines::PipeOp$predict()
• mlr3pipelines::PipeOp$print()
• mlr3pipelines::PipeOp$train()
• mlr3torch::PipeOpTorch$shapes_out()

---

Method new()

Creates a new instance of this R6 class.

Usage

PipeOpTorchFn$new(
  fn,
  id = "nn_fn",
  param_vals = list(),
  param_set = NULL,
  shapes_out = NULL
)

Arguments

fn

(function)
The function to be applied. Takes a torch tensor as first argument and returns a torch tensor.

id

(character(1))
Identifier of the resulting object.

param_vals

(list())
List of hyperparameter settings, overwriting the hyperparameter settings that would otherwise be set during construction.

param_set

(ParamSet or NULL)
A ParamSet wrapping the arguments to fn. If omitted, then the ParamSet for this PipeOp will be inferred from the function signature.

shapes_out

(function or NULL)
A function that computes the output shapes of the fn. See PipeOpTorch's .shapes_out() method for details on the parameters, and PipeOpTaskPreprocTorch for details on how the shapes are inferred when this parameter is NULL.

---

Method clone()

The objects of this class are cloneable with this method.

Usage

PipeOpTorchFn$clone(deep = FALSE)

Arguments

deep

Whether to make a deep clone.

Examples

custom_fn =  function(x, a) x / a
obj = po("nn_fn", fn = custom_fn, a = 2)
obj$param_set
#> <ParamSet(1)>
#>        id    class lower upper nlevels        default  value
#>    <char>   <char> <num> <num>   <num>         <list> <list>
#> 1:      a ParamUty    NA    NA     Inf <NoDefault[0]>      2

graph = po("torch_ingress_ltnsr") %>>% obj

task = tsk("lazy_iris")$filter(1)
tnsr = materialize(task$data()$x)[[1]]

md_trained = graph$train(task)
trained = md_trained[[1]]$graph$train(tnsr)

trained[[1]]
#> torch_tensor
#>  2.5500
#>  1.7500
#>  0.7000
#>  0.1000
#> [ CPUFloatType{4} ]

custom_fn(tnsr, a = 2)
#> torch_tensor
#>  2.5500
#>  1.7500
#>  0.7000
#>  0.1000
#> [ CPUFloatType{4} ]