Compile Keras Models Over a Grid of Hyperparameters

Description

Pre-compiles Keras models for each hyperparameter combination in a grid.

This function is a powerful debugging tool to use before running a full tune::tune_grid(). It allows you to quickly validate multiple model architectures, ensuring they can be successfully built and compiled without the time-consuming process of actually fitting them. It helps catch common errors like incompatible layer shapes or invalid argument values early.

Usage

compile_keras_grid(spec, grid, x, y)

Arguments

Details

Compile and Validate Keras Model Architectures

The function iterates through each row of the provided grid. For each hyperparameter combination, it attempts to build and compile the Keras model defined by the spec. The process is wrapped in a try-catch block to gracefully handle and report any errors that occur during model instantiation or compilation.

The output is a tibble that mirrors the input grid, with additional columns containing the compiled model object or the error message, making it easy to inspect which architectures are valid.

Value

A tibble with the following columns:

• Columns from the input grid.

• compiled_model: A list-column containing the compiled Keras model objects. If compilation failed, the element will be NULL.

• error: A list-column containing NA for successes or a character string with the error message for failures.

Examples

if (requireNamespace("keras3", quietly = TRUE)) {
library(keras3)
library(parsnip)
library(dials)

# 1. Define layer blocks
input_block <- function(model, input_shape) {
  keras_model_sequential(input_shape = input_shape)
}
hidden_block <- function(model, units = 32) {
  model |> layer_dense(units = units, activation = "relu")
}
output_block <- function(model, num_classes) {
  model |> layer_dense(units = num_classes, activation = "softmax")
}

# 2. Define a kerasnip model specification
create_keras_sequential_spec(
  model_name = "my_mlp_grid",
  layer_blocks = list(
    input = input_block,
    hidden = hidden_block,
    output = output_block
  ),
  mode = "classification"
)

mlp_spec <- my_mlp_grid(
  hidden_units = tune(),
  compile_loss = "categorical_crossentropy",
  compile_optimizer = "adam"
)

# 3. Create a hyperparameter grid
# Include an invalid value (-10) to demonstrate error handling
param_grid <- tibble::tibble(
  hidden_units = c(32, 64, -10)
)

# 4. Prepare dummy data
x_train <- matrix(rnorm(100 * 10), ncol = 10)
y_train <- factor(sample(0:1, 100, replace = TRUE))

# 5. Compile models over the grid
compiled_grid <- compile_keras_grid(
  spec = mlp_spec,
  grid = param_grid,
  x = x_train,
  y = y_train
)

print(compiled_grid)
remove_keras_spec("my_mlp_grid")

# 6. Inspect the results
# The row with `hidden_units = -10` will show an error.
}

Create a Custom Keras Functional API Model Specification for Tidymodels

Description

This function acts as a factory to generate a new parsnip model specification based on user-defined blocks of Keras layers using the Functional API. This allows for creating complex, tunable architectures with non-linear topologies that integrate seamlessly with the tidymodels ecosystem.

Usage

create_keras_functional_spec(
  model_name,
  layer_blocks,
  mode = c("regression", "classification"),
  ...,
  env = parent.frame()
)

Arguments

Details

This function generates all the boilerplate needed to create a custom, tunable parsnip model specification that uses the Keras Functional API. This is ideal for models with complex, non-linear topologies, such as networks with multiple inputs/outputs or residual connections.

The function inspects the arguments of your layer_blocks functions and makes them available as tunable parameters in the generated model specification, prefixed with the block's name (e.g., dense_units). Common training parameters such as epochs and learn_rate are also added.

Value

Invisibly returns NULL. Its primary side effect is to create a new model specification function (e.g., custom_resnet()) in the specified environment and register the model with parsnip so it can be used within the tidymodels framework.

Model Graph Connectivity

kerasnip builds the model's directed acyclic graph by inspecting the arguments of each function in the layer_blocks list. The connection logic is as follows:

1. The names of the elements in the layer_blocks list define the names of the nodes in your graph (e.g., main_input, dense_path, output).

2. The names of the arguments in each block function specify its inputs. A block function like ⁠my_block <- function(input_a, input_b, ...)⁠ declares that it needs input from the nodes named input_a and input_b. kerasnip will automatically supply the output tensors from those nodes when calling my_block.

There are two special requirements:

• Input Block: The first block in the list is treated as the input node. Its function should not take other blocks as input, but it can have an input_shape argument, which is supplied automatically during fitting.

• Output Block: Exactly one block must be named "output". The tensor returned by this block is used as the final output of the Keras model.

A key feature is the automatic creation of ⁠num_{block_name}⁠ arguments (e.g., num_dense_path). This allows you to control how many times a block is repeated, making it easy to tune the depth of your network. A block can only be repeated if it has exactly one input from another block in the graph.

The new model specification function and its update() method are created in the environment specified by the env argument.

See Also

remove_keras_spec(), parsnip::new_model_spec(), create_keras_sequential_spec()

Examples

if (requireNamespace("keras3", quietly = TRUE)) {
  library(keras3)
  library(parsnip)

  # 1. Define block functions. These are the building blocks of our model.
  # An input block that receives the data's shape automatically.
  input_block <- function(input_shape) layer_input(shape = input_shape)

  # A dense block with a tunable `units` parameter.
  dense_block <- function(tensor, units) {
    tensor |> layer_dense(units = units, activation = "relu")
  }

  # A block that adds two tensors together (for the residual connection).
  add_block <- function(input_a, input_b) layer_add(list(input_a, input_b))

  # An output block for regression.
  output_block_reg <- function(tensor) layer_dense(tensor, units = 1)

  # 2. Create the spec. The `layer_blocks` list defines the graph.
  create_keras_functional_spec(
    model_name = "my_resnet_spec",
    layer_blocks = list(
      # The names of list elements are the node names.
      main_input = input_block,

      # The argument `main_input` connects this block to the input node.
      dense_path = function(main_input, units = 32) dense_block(main_input, units),

      # This block's arguments connect it to the original input AND the dense layer.
      add_residual = function(main_input, dense_path) add_block(main_input, dense_path),

      # This block must be named 'output'. It connects to the residual add layer.
      output = function(add_residual) output_block_reg(add_residual)
    ),
    mode = "regression"
  )

  # 3. Use the newly created specification function!
  # The `dense_path_units` argument was created automatically.
  model_spec <- my_resnet_spec(dense_path_units = 64, epochs = 10)

  # You could also tune the number of dense layers since it has a single input:
  # model_spec <- my_resnet_spec(num_dense_path = 2, dense_path_units = 32)

  print(model_spec)
  remove_keras_spec("my_resnet_spec")
  # tune::tunable(model_spec)
}

Create a Custom Keras Sequential Model Specification for Tidymodels

Description

This function acts as a factory to generate a new parsnip model specification based on user-defined blocks of Keras layers using the Sequential API. This is the ideal choice for creating models that are a simple, linear stack of layers. For models with complex, non-linear topologies, see create_keras_functional_spec().

Usage

create_keras_sequential_spec(
  model_name,
  layer_blocks,
  mode = c("regression", "classification"),
  ...,
  env = parent.frame()
)

Arguments

Details

This function generates all the boilerplate needed to create a custom, tunable parsnip model specification that uses the Keras Sequential API.

The function inspects the arguments of your layer_blocks functions (ignoring special arguments like input_shape and num_classes) and makes them available as arguments in the generated model specification, prefixed with the block's name (e.g., dense_units).

The new model specification function and its update() method are created in the environment specified by the env argument.

Value

Invisibly returns NULL. Its primary side effect is to create a new model specification function (e.g., my_mlp()) in the specified environment and register the model with parsnip so it can be used within the tidymodels framework.

Model Architecture (Sequential API)

kerasnip builds the model by applying the functions in layer_blocks in the order they are provided. Each function receives the Keras model built by the previous function and returns a modified version.

1. The first block must initialize the model (e.g., with keras_model_sequential()). It can accept an input_shape argument, which kerasnip will provide automatically during fitting.

2. Subsequent blocks add layers to the model.

3. The final block should add the output layer. For classification, it can accept a num_classes argument, which is provided automatically.

A key feature of this function is the automatic creation of ⁠num_{block_name}⁠ arguments (e.g., num_hidden). This allows you to control how many times each block is repeated, making it easy to tune the depth of your network.

See Also

remove_keras_spec(), parsnip::new_model_spec(), create_keras_functional_spec()

Examples

if (requireNamespace("keras3", quietly = TRUE)) {
library(keras3)
library(parsnip)
library(dials)

# 1. Define layer blocks for a complete model.
# The first block must initialize the model. `input_shape` is passed automatically.
input_block <- function(model, input_shape) {
  keras_model_sequential(input_shape = input_shape)
}
# A block for hidden layers. `units` will become a tunable parameter.
hidden_block <- function(model, units = 32) {
  model |> layer_dense(units = units, activation = "relu")
}

# The output block. `num_classes` is passed automatically for classification.
output_block <- function(model, num_classes) {
  model |> layer_dense(units = num_classes, activation = "softmax")
}

# 2. Create the spec, providing blocks in the correct order.
create_keras_sequential_spec(
model_name = "my_mlp_seq_spec",
  layer_blocks = list(
    input = input_block,
    hidden = hidden_block,
    output = output_block
  ),
  mode = "classification"
)

# 3. Use the newly created specification function!
# Note the new arguments `num_hidden` and `hidden_units`.
model_spec <- my_mlp_seq_spec(
  num_hidden = 2,
  hidden_units = 64,
  epochs = 10,
  learn_rate = 0.01
)

print(model_spec)
remove_keras_spec("my_mlp_seq_spec")
}

Extract Keras Training History

Description

Extracts and returns the training history from a parsnip model_fit object created by kerasnip.

Usage

extract_keras_history(object)

Arguments

Details

Extract Keras Training History

The history object contains the metrics recorded during model training, such as loss and accuracy, for each epoch. This is highly useful for visualizing the training process and diagnosing issues like overfitting. The returned object can be plotted directly.

Value

A keras_training_history object. You can call plot() on this object to visualize the learning curves.

See Also

keras_evaluate, extract_keras_model

Extract Keras Model from a Fitted Kerasnip Object

Description

Extracts and returns the underlying Keras model object from a parsnip model_fit object created by kerasnip.

Usage

extract_keras_model(object)

Arguments

Details

Extract the Raw Keras Model from a Kerasnip Fit

This is useful when you need to work directly with the Keras model object for tasks like inspecting layer weights, creating custom plots, or passing it to other Keras-specific functions.

Value

The raw Keras model object (keras_model).

See Also

keras_evaluate, extract_keras_history

Extract Valid Grid from Compilation Results

Description

This helper function filters the results from compile_keras_grid() to return a new hyperparameter grid containing only the combinations that compiled successfully.

Usage

extract_valid_grid(compiled_grid)

Arguments

Details

Filter a Grid to Only Valid Hyperparameter Sets

After running compile_keras_grid(), you can use this function to remove problematic hyperparameter combinations before proceeding to the full tune::tune_grid().

Value

A tibble containing the subset of the original grid that resulted in a successful model compilation. The compiled_model and error columns are removed, leaving a clean grid ready for tuning.

Examples

if (requireNamespace("keras3", quietly = TRUE)) {
  library(keras3)
  library(parsnip)
  library(dials)

  # 1. Define layer blocks
  input_block <- function(model, input_shape) {
    keras_model_sequential(input_shape = input_shape)
  }
  hidden_block <- function(model, units = 32) {
    model |> layer_dense(units = units, activation = "relu")
  }
  output_block <- function(model, num_classes) {
    model |> layer_dense(units = num_classes, activation = "softmax")
  }

  # 2. Define a kerasnip model specification
  create_keras_sequential_spec(
    model_name = "my_mlp_grid_2",
    layer_blocks = list(
      input = input_block,
      hidden = hidden_block,
      output = output_block
    ),
    mode = "classification"
  )

  mlp_spec <- my_mlp_grid_2(
    hidden_units = tune(),
    compile_loss = "categorical_crossentropy",
    compile_optimizer = "adam"
  )

  # 3. Create a hyperparameter grid
  param_grid <- tibble::tibble(
    hidden_units = c(32, 64, -10)
  )

  # 4. Prepare dummy data
  x_train <- matrix(rnorm(100 * 10), ncol = 10)
  y_train <- factor(sample(0:1, 100, replace = TRUE))

  # 5. Compile models over the grid
  compiled_grid <- compile_keras_grid(
    spec = mlp_spec,
    grid = param_grid,
    x = x_train,
    y = y_train
  )

  # 6. Extract the valid grid
  valid_grid <- extract_valid_grid(compiled_grid)
  print(valid_grid)
  remove_keras_spec("my_mlp_grid_2")
}

Internal Fitting Engine for Functional API Models

Description

This function serves as the internal engine for fitting kerasnip models that are based on the Keras functional API. It is not intended to be called directly by the user. The function is invoked by parsnip::fit() when a kerasnip functional model specification is used.

Usage

generic_functional_fit(formula, data, layer_blocks, ...)

Arguments

Details

Generic Fitting Function for Functional Keras Models

The function orchestrates the three main steps of the model fitting process:

1. Build and Compile: It calls build_and_compile_functional_model() to construct the Keras model architecture based on the provided layer_blocks and hyperparameters.

2. Process Data: It preprocesses the input (x) and output (y) data into the format expected by Keras.

3. Fit Model: It calls keras3::fit() with the compiled model and processed data, passing along any fitting-specific arguments (e.g., epochs, batch_size, callbacks).

Value

A list containing the fitted model and other metadata. This list is stored in the fit slot of the parsnip model fit object. The list contains the following elements:

• fit: The raw, fitted Keras model object.

• history: The Keras training history object.

• lvl: A character vector of the outcome factor levels (for classification) or NULL (for regression).

Examples

# This function is not called directly by users.
# It is called internally by `parsnip::fit()`.
# For example:

library(parsnip)
# create_keras_functional_spec(...) defines my_functional_model

# spec <- my_functional_model(hidden_units = 128, fit_epochs = 10) |>
#   set_engine("keras")

# # This call to fit() would invoke generic_functional_fit() internally
# fitted_model <- fit(spec, y ~ x, data = training_data)

Internal Fitting Engine for Sequential API Models

Description

This function serves as the internal engine for fitting kerasnip models that are based on the Keras sequential API. It is not intended to be called directly by the user. The function is invoked by parsnip::fit() when a kerasnip sequential model specification is used.

Usage

generic_sequential_fit(formula, data, layer_blocks, ...)

Arguments

Details

Generic Fitting Function for Sequential Keras Models

The function orchestrates the three main steps of the model fitting process:

1. Build and Compile: It calls build_and_compile_sequential_model() to construct the Keras model architecture based on the provided layer_blocks and hyperparameters.

2. Process Data: It preprocesses the input (x) and output (y) data into the format expected by Keras.

3. Fit Model: It calls keras3::fit() with the compiled model and processed data, passing along any fitting-specific arguments (e.g., epochs, batch_size, callbacks).

Value

A list containing the fitted model and other metadata. This list is stored in the fit slot of the parsnip model fit object. The list contains the following elements:

• fit: The raw, fitted Keras model object.

• history: The Keras training history object.

• lvl: A character vector of the outcome factor levels (for classification) or NULL (for regression).

Examples

# This function is not called directly by users.
# It is called internally by `parsnip::fit()`.
# For example:

library(parsnip)
# create_keras_sequential_spec(...) defines my_sequential_model

# spec <- my_sequential_model(hidden_1_units = 128, fit_epochs = 10) |>
#   set_engine("keras")

# # This call to fit() would invoke generic_sequential_fit() internally
# fitted_model <- fit(spec, y ~ x, data = training_data)

Get Parsnip's Model Environment

Description

This is an internal helper function to retrieve the environment where parsnip stores its model definitions. It is used to dynamically interact with the parsnip infrastructure.

Usage

get_model_env()

Value

The parsnip model environment.

Examples

model_env <- kerasnip::get_model_env()

Inform About Compilation Errors

Description

This helper function inspects the results from compile_keras_grid() and prints a formatted, easy-to-read summary of any compilation errors that occurred.

Usage

inform_errors(compiled_grid, n = 10)

Arguments

Details

Display a Summary of Compilation Errors

This is most useful for interactive debugging of complex tuning grids where some hyperparameter combinations may lead to invalid Keras models.

Value

Invisibly returns the input compiled_grid. Called for its side effect of printing a summary to the console.

Examples

if (requireNamespace("keras3", quietly = TRUE)) {
  library(keras3)
  library(parsnip)
  library(dials)

  # 1. Define layer blocks
  input_block <- function(model, input_shape) {
    keras_model_sequential(input_shape = input_shape)
  }
  hidden_block <- function(model, units = 32) {
    model |> layer_dense(units = units, activation = "relu")
  }
  output_block <- function(model, num_classes) {
    model |> layer_dense(units = num_classes, activation = "softmax")
  }

  # 2. Define a kerasnip model specification
  create_keras_sequential_spec(
    model_name = "my_mlp_grid_3",
    layer_blocks = list(
      input = input_block,
      hidden = hidden_block,
      output = output_block
    ),
    mode = "classification"
  )

  mlp_spec <- my_mlp_grid_3(
    hidden_units = tune(),
    compile_loss = "categorical_crossentropy",
    compile_optimizer = "adam"
  )

  # 3. Create a hyperparameter grid
  param_grid <- tibble::tibble(
    hidden_units = c(32, 64, -10)
  )

  # 4. Prepare dummy data
  x_train <- matrix(rnorm(100 * 10), ncol = 10)
  y_train <- factor(sample(0:1, 100, replace = TRUE))

  # 5. Compile models over the grid
  compiled_grid <- compile_keras_grid(
    spec = mlp_spec,
    grid = param_grid,
    x = x_train,
    y = y_train
  )

  # 6. Inform about errors
  inform_errors(compiled_grid)
  remove_keras_spec("my_mlp_grid_3")
}

Remap Layer Block Arguments for Model Specification

Description

Creates a wrapper function around a Keras layer block to rename its arguments. This is a powerful helper for defining the layer_blocks in create_keras_functional_spec() and create_keras_sequential_spec(), allowing you to connect reusable blocks into a model graph without writing verbose anonymous functions.

Usage

inp_spec(block, input_map)

Arguments

Details

inp_spec() makes your model definitions cleaner and more readable. It handles the metaprogramming required to create a new function with the correct argument names, while preserving the original block's hyperparameters and their default values.

The function supports two modes of operation based on input_map:

1. Single Input Renaming: If input_map is a single character string, the wrapper function renames the first argument of the block function to the provided string. This is the common case for blocks that take a single tensor input.

2. Multiple Input Mapping: If input_map is a named character vector, it provides an explicit mapping from new argument names (the names of the vector) to the original argument names in the block function (the values of the vector). This is used for blocks with multiple inputs, like a concatenation layer.

Value

A new function (a closure) that wraps the block function with renamed arguments, ready to be used in a layer_blocks list.

Examples

# --- Example Blocks ---
# A standard dense block with one input tensor and one hyperparameter.
dense_block <- function(tensor, units = 16) {
  tensor |> keras3::layer_dense(units = units, activation = "relu")
}

# A block that takes two tensors as input.
concat_block <- function(input_a, input_b) {
  keras3::layer_concatenate(list(input_a, input_b))
}

# An output block with one input.
output_block <- function(tensor) {
  tensor |> keras3::layer_dense(units = 1)
}

# --- Usage ---
layer_blocks <- list(
  main_input = keras3::layer_input,
  path_a = inp_spec(dense_block, "main_input"),
  path_b = inp_spec(dense_block, "main_input"),
  concatenated = inp_spec(
    concat_block,
    c(path_a = "input_a", path_b = "input_b")
  ),
  output = inp_spec(output_block, "concatenated")
)

Evaluate a Kerasnip Model

Description

This function provides an kera_evaluate() method for model_fit objects created by kerasnip. It preprocesses the new data into the format expected by Keras and then calls keras3::evaluate() on the underlying model to compute the loss and any other metrics.

Usage

keras_evaluate(object, x, y = NULL, ...)

Arguments

Details

Evaluate a Fitted Kerasnip Model on New Data

Value

A named list containing the evaluation results (e.g., loss, accuracy). The names are determined by the metrics the model was compiled with.

Examples

if (requireNamespace("keras3", quietly = TRUE)) {
library(keras3)
library(parsnip)

# 1. Define layer blocks
input_block <- function(model, input_shape) {
  keras_model_sequential(input_shape = input_shape)
}
hidden_block <- function(model, units = 32) {
  model |> layer_dense(units = units, activation = "relu")
}
output_block <- function(model, num_classes) {
  model |> layer_dense(units = num_classes, activation = "softmax")
}

# 2. Define and fit a model ----
create_keras_sequential_spec(
  model_name = "my_mlp_tools",
  layer_blocks = list(
    input = input_block,
    hidden = hidden_block,
    output = output_block
  ),
  mode = "classification"
)

mlp_spec <- my_mlp_tools(
  hidden_units = 32,
  compile_loss = "categorical_crossentropy",
  compile_optimizer = "adam",
  compile_metrics = "accuracy",
  fit_epochs = 5
) |> set_engine("keras")

x_train <- matrix(rnorm(100 * 10), ncol = 10)
y_train <- factor(sample(0:1, 100, replace = TRUE))
train_df <- data.frame(x = I(x_train), y = y_train)

fitted_mlp <- fit(mlp_spec, y ~ x, data = train_df)

# 3. Evaluate the model on new data ----
x_test <- matrix(rnorm(50 * 10), ncol = 10)
y_test <- factor(sample(0:1, 50, replace = TRUE))

eval_metrics <- keras_evaluate(fitted_mlp, x_test, y_test)
print(eval_metrics)

# 4. Extract the Keras model object ----
keras_model <- extract_keras_model(fitted_mlp)
summary(keras_model)

# 5. Extract the training history ----
history <- extract_keras_history(fitted_mlp)
plot(history)
remove_keras_spec("my_mlp_tools")
}

Dynamically Discovered Keras Objects

Description

These exported vectors contain the names of optimizers, losses, and metrics discovered from the installed keras3 package when kerasnip is loaded. This ensures that kerasnip is always up-to-date with your Keras version.

Usage

keras_optimizers

keras_losses

keras_metrics

Format

An object of class character of length 12.

An object of class character of length 21.

An object of class character of length 32.

Details

These objects are primarily used to provide the default values for the dials parameter functions, optimizer_function() and loss_function_keras(). This allows for tab-completion in IDEs and validation of optimizer and loss names when tuning models.

The discovery process in .onLoad() scrapes the keras3 namespace for functions matching ⁠optimizer_*⁠, ⁠loss_*⁠, and ⁠metric_*⁠ patterns.

Dials Parameter for Keras Loss Functions

Description

Dials Parameter for Keras Loss Functions

Usage

loss_function_keras(values = NULL)

Arguments

Value

A dials parameter object for Keras loss.

Check if a Kerasnip Model Specification Exists

Description

This is an internal helper function to check if a model specification has been registered in the parsnip model environment.

Usage

model_exists(model_name)

Arguments

Value

A logical value, TRUE if the model exists, FALSE otherwise.

Examples

if (requireNamespace("parsnip", quietly = TRUE)) {
  library(parsnip)

  # Check for a model that exists in parsnip
  model_exists("mlp")

  # Check for a model that does not exist
  model_exists("non_existent_model")
}

Dials Parameter for Keras Optimizers

Description

Dials Parameter for Keras Optimizers

Usage

optimizer_function(values = NULL)

Arguments

Value

A dials parameter object for Keras optimizers.

Process Predictor Input for Keras (Functional API)

Description

Preprocesses predictor data (x) into a format suitable for Keras models built with the Functional API. Handles both tabular data and list-columns of arrays (e.g., for images), supporting multiple inputs.

Usage

process_x_functional(x)

Arguments

Value

A list containing:

• x_proc: The processed predictor data (matrix or array, or list of arrays).

• input_shape: The determined input shape(s) for the Keras model.

Process Predictor Input for Keras

Description

Preprocesses predictor data (x) into a format suitable for Keras models. Handles both tabular data and list-columns of arrays (e.g., for images).

Usage

process_x_sequential(x)

Arguments

Value

A list containing:

• x_proc: The processed predictor data (matrix or array).

• input_shape: The determined input shape for the Keras model.

Process Outcome Input for Keras (Functional API)

Description

Preprocesses outcome data (y) into a format suitable for Keras models built with the Functional API. Handles both regression (numeric) and classification (factor) outcomes, including one-hot encoding for classification, and supports multiple outputs.

Usage

process_y_functional(y, is_classification = NULL, class_levels = NULL)

Arguments

Value

A list containing:

• y_proc: The processed outcome data (matrix or one-hot encoded array, or list of these for multiple outputs).

• is_classification: Logical, indicating if y was treated as classification.

• num_classes: Integer, the number of classes for classification, or NULL.

• class_levels: Character vector, the factor levels for classification, or NULL.

Process Outcome Input for Keras

Description

Preprocesses outcome data (y) into a format suitable for Keras models. Handles both regression (numeric) and classification (factor) outcomes, including one-hot encoding for classification.

Usage

process_y_sequential(y, is_classification = NULL, class_levels = NULL)

Arguments

Value

A list containing:

• y_proc: The processed outcome data (matrix or one-hot encoded array).

• is_classification: Logical, indicating if y was treated as classification.

• num_classes: Integer, the number of classes for classification, or NULL.

• class_levels: Character vector, the factor levels for classification, or NULL.

Register a Custom Keras Loss

Description

Allows users to register a custom loss function so it can be used by name within kerasnip model specifications and tuned with dials.

Usage

register_keras_loss(name, loss_fn)

Arguments

Details

Registered losses are stored in an internal environment. When a model is compiled, kerasnip will first check this internal registry for a loss matching the provided name before checking the keras3 package.

Value

No return value, called for side effects.

See Also

register_keras_optimizer(), register_keras_metric()

Register a Custom Keras Metric

Description

Allows users to register a custom metric function so it can be used by name within kerasnip model specifications.

Usage

register_keras_metric(name, metric_fn)

Arguments

Details

Registered metrics are stored in an internal environment. When a model is compiled, kerasnip will first check this internal registry for a metric matching the provided name before checking the keras3 package.

Value

No return value, called for side effects.

See Also

register_keras_optimizer(), register_keras_loss()

Register a Custom Keras Optimizer

Description

Allows users to register a custom optimizer function so it can be used by name within kerasnip model specifications and tuned with dials.

Usage

register_keras_optimizer(name, optimizer_fn)

Arguments

Details

Registered optimizers are stored in an internal environment. When a model is compiled, kerasnip will first check this internal registry for an optimizer matching the provided name before checking the keras3 package.

The optimizer_fn can be a simple function or a partially applied function using purrr::partial(). This is useful for creating versions of Keras optimizers with specific settings.

Value

No return value, called for side effects.

See Also

register_keras_loss(), register_keras_metric()

Examples

if (requireNamespace("keras3", quietly = TRUE)) {
  # Register a custom version of Adam with a different default beta_1
  my_adam <- purrr::partial(keras3::optimizer_adam, beta_1 = 0.8)
  register_keras_optimizer("my_adam", my_adam)

  # Now "my_adam" can be used as a string in a model spec, e.g.,
  # my_model_spec(compile_optimizer = "my_adam")
}

Remove a Keras Model Specification and its Registrations

Description

This function completely removes a model specification that was previously created by create_keras_sequential_spec() or create_keras_functional_spec(). It cleans up both the function in the user's environment and all associated registrations within the parsnip package.

Usage

remove_keras_spec(model_name, env = parent.frame())

Arguments

Details

This function is essential for cleanly unloading a dynamically created model. It performs three main actions:

1. It removes the model specification function (e.g., my_mlp()) and its corresponding update() method from the specified environment.

2. It searches parsnip's internal model environment for all objects whose names start with the model_name and removes them. This purges the fit methods, argument definitions, and other registrations.

3. It removes the model's name from parsnip's master list of models.

This function uses the un-exported get_model_env() to perform the cleanup.

Value

Invisibly returns TRUE after attempting to remove the objects.

See Also

create_keras_sequential_spec(), create_keras_functional_spec()

Examples

if (requireNamespace("keras3", quietly = TRUE)) {
  # First, create a dummy spec
  input_block <- function(model, input_shape) {
    keras3::keras_model_sequential(input_shape = input_shape)
  }
  dense_block <- function(model, units = 16) {
    model |> keras3::layer_dense(units = units)
  }
  create_keras_sequential_spec(
    "my_temp_model",
    list(
      input = input_block,
      dense = dense_block
    ),
    "regression"
  )

  # Check it exists in the environment and in parsnip
  exists("my_temp_model")
  "my_temp_model" %in% parsnip::show_engines("my_temp_model")$model

  # Now remove it
  remove_keras_spec("my_temp_model")

  # Check it's gone
  !exists("my_temp_model")
  !model_exists("my_temp_model")
}

Collapse Predictors into a single list-column

Description

step_collapse() creates a a specification of a recipe step that will convert a group of predictors into a single list-column. This is useful for custom models that need the predictors in a different format.

Usage

step_collapse(
  recipe,
  ...,
  role = "predictor",
  trained = FALSE,
  columns = NULL,
  new_col = "predictor_matrix",
  skip = FALSE,
  id = recipes::rand_id("collapse")
)

Arguments

Value

An updated version of recipe with the new step added to the sequence of existing steps (if any). For the tidy method, a tibble with columns terms which is the columns that are affected and value which is the type of collapse.

Examples

library(recipes)

# 2 predictors
dat <- data.frame(
  x1 = 1:10,
  x2 = 11:20,
  y = 1:10
)

rec <- recipe(y ~ ., data = dat) %>%
  step_collapse(x1, x2, new_col = "pred") %>%
  prep()

bake(rec, new_data = NULL)