new_act_fn

<a href="https://lifecycle.r-lib.org/articles/stages.html#experimental"><img src="figures/lifecycle-experimental.svg?package=kindling&version=0.3.0" alt="[Experimental]"></a>
Wraps a user-supplied function into a validated custom activation,
ensuring it accepts and returns a <code>torch_tensor</code>. Performs an eager
dry-run probe at definition time so errors surface early, and
wraps the function with a call-time type guard for safety.

Provides a higher-level interface to the 'torch' package for defining,
training, and fine-tuning neural networks, including its depth, powered by code generation.
This package supports few to several architectures, including feedforward (multi-layer perceptron)
and recurrent neural networks (Recurrent Neural Networks (RNN), Long Short-Term Memory (LSTM), Gated Recurrent Unit (GRU)),
while also reduces boilerplate 'torch' code while enabling seamless integration with 'torch'. The model methods
to train neural networks from this package also bridges to titanic ML frameworks in R, namely
'tidymodels' ecosystem, which enables the 'parsnip' model specifications, workflows, recipes,
and tuning tools.

Joshua Marie

kindling

Higher-Level Interface of 'torch' Package to Auto-Train Neural
Networks

Antoine Soetewey

new_act_fn function

<dl><dt>fn</dt>
<dd>A function taking a single tensor argument and returning a tensor.
E.g. <code>\(x) torch::torch_tanh(x)</code>.</dd>
<dt>probe</dt>
<dd>Logical. If <code>TRUE</code> (default), runs a dry-run with a small
dummy tensor at definition time to catch obvious errors early.</dd>
<dt>.name</dt>
<dd>A string to be stored in an attribute. Nothing special, except it is
used when displaying the info of a trained neural network model.
Default is <code>"&lt;custom&gt;"</code>.</dd></dl>

Arguments

<a href='https://lifecycle.r-lib.org/articles/stages.html#experimental'><img src='figures/lifecycle-experimental.svg' alt='[Experimental]' /></a>
Wraps a user-supplied function into a validated custom activation,
ensuring it accepts and returns a <code>torch_tensor</code>. Performs an eager
dry-run probe at definition time so errors surface early, and
wraps the function with a call-time type guard for safety.

Custom Activation Function Constructor — new_act_fn

<dl>

<dt>fn</dt>
<dd>A function taking a single tensor argument and returning a tensor.
E.g. <code>\(x) torch::torch_tanh(x)</code>.</dd>


<dt>probe</dt>
<dd>Logical. If <code>TRUE</code> (default), runs a dry-run with a small
dummy tensor at definition time to catch obvious errors early.</dd>


<dt>.name</dt>
<dd>A string to be stored in an attribute. Nothing special, except it is
used when displaying the info of a trained neural network model.
Default is <code>"&lt;custom&gt;"</code>.</dd>

</dl>

new_act_fn: Custom Activation Function Constructor

Description

Usage

Value

Arguments