LearnerMultioutputKeras: Keras Neural Network with custom architecture (Multilabel)

In mlr-org/mlr3multioutput: Multioutput extension for 'mlr3'

Description

Neural Network using Keras and Tensorflow. This learner allows for supplying a custom architecture. Calls keras::fit() from package keras.

Parameters: Most of the parameters can be obtained from the keras documentation. Some exceptions are documented here.

• model: A compiled keras model suited for the task.

• class_weight: A named list of class-weights for the different classes numbered from 0 to c-1 (for c classes).

  Example:
  wts = c(0.5, 1)
  setNames(as.list(wts), seq_len(length(wts)) - 1)
  

• callbacks: A list of keras callbacks. See ?callbacks.

Format

R6::R6Class() inheriting from mlr3::LearnerMultioutput.

Construction

LearnerMultioutputKeras$new()
mlr3::mlr_learners$get("multiout.keras")
mlr3::lrn("multiout.keras")

Learner Methods

Keras Learners offer several methods for easy access to the stored models.

• .$plot()
  Plots the history, i.e. the train-validation loss during training.

• .$save(file_path)
  Dumps the model to a provided file_path in 'h5' format.

• .$load_model_from_file(file_path)
  Loads a model saved using saved back into the learner. The model needs to be saved separately when the learner is serialized. In this case, the learner can be restored from this function. Currently not implemented for 'TabNet'.

• .$lr_find(task, epochs, lr_min, lr_max, batch_size)
  Employ an implementation of the learning rate finder as popularized by Jeremy Howard in fast.ai (http://course.fast.ai/) for the learner. For more info on parameters, see find_lr.

Super classes

mlr3::Learner -> mlr3multioutput::LearnerMultioutput -> LearnerMultioutputKeras

Methods

Public methods

• LearnerMultioutputKeras$new()

• LearnerMultioutputKeras$save()

• LearnerMultioutputKeras$load_model_from_file()

• LearnerMultioutputKeras$plot()

• LearnerMultioutputKeras$lr_find()

• LearnerMultioutputKeras$clone()

Inherited methods

• mlr3::Learner$format()

• mlr3::Learner$help()

• mlr3::Learner$predict()

• mlr3::Learner$predict_newdata()

• mlr3::Learner$print()

• mlr3::Learner$reset()

• mlr3::Learner$train()

---

Method new()

Usage

LearnerMultioutputKeras$new(
  id = "multiout.keras",
  predict_types = c("response", "prob"),
  feature_types = c("integer", "numeric"),
  properties = c("multilabel"),
  packages = "keras",
  man = "mlr3keras::mlr_learners_classif.keras",
  architecture = KerasArchitectureCustomModel$new()
)

---

Method save()

Usage

LearnerMultioutputKeras$save(filepath)

---

Method load_model_from_file()

Usage

LearnerMultioutputKeras$load_model_from_file(filepath)

---

Method plot()

Usage

LearnerMultioutputKeras$plot()

---

Method lr_find()

Usage

LearnerMultioutputKeras$lr_find(
  task,
  epochs = 5L,
  lr_min = 10^-4,
  lr_max = 0.8,
  batch_size = 128L
)

---

Method clone()

The objects of this class are cloneable with this method.

Usage

LearnerMultioutputKeras$clone(deep = FALSE)

Arguments

deep

Whether to make a deep clone.

See Also

Dictionary of Learners: mlr3::mlr_learners

Examples

 # Define a model
 library(keras)
 model = keras_model_sequential() %>%
 layer_dense(units = 12L, input_shape = 4L, activation = "relu") %>%
 layer_dense(units = 12L, activation = "relu") %>%
 layer_dense(units = 3L, activation = "sigmoid") %>%
   compile(optimizer = optimizer_sgd(),
     loss = "binary_crossentropy",
     metrics = "accuracy")
 # Create the learner
 learner = LearnerMultioutputKeras$new()
 learner$param_set$values$model = model
 learner$train(mlr3::mlr_tasks$get("flags"))

mlr-org/mlr3multioutput documentation built on Nov. 22, 2020, 1:17 p.m.