tests/testthat/test_classif_kerasff.R

In mlr-org/mlr3keras: mlr3 Keras extension

test_that("autotest", {
  skip_on_os("solaris")
  learner = mlr3::lrn("classif.kerasff")
  expect_learner(learner)

  learner$param_set$values$epochs = 3L
  result = run_autotest(learner, exclude = "(feat_single|sanity)", check_replicable = FALSE)
  expect_true(result, info = result$error)
  k_clear_session()
})

test_that("can fit logistic regression", {
  skip_on_os("solaris")
  learner = mlr3::lrn("classif.kerasff")
  expect_learner(learner)

  learner$param_set$values$epochs = 3L
  learner$param_set$values$layer_units = integer()
  learner$train(mlr_tasks$get("iris"))

  expect_list(learner$state)
  expect_list(learner$state$model)
  prd = learner$predict(mlr_tasks$get("iris"))
  expect_class(prd, "PredictionClassif")
  k_clear_session()
})

test_that("works with pipelines", {
  skip_on_os("solaris")
  skip_if_not(require("mlr3pipelines"))

  po_imp = PipeOpImputeMedian$new()
  po_lrn = PipeOpLearner$new(lrn("classif.kerasff"))
  po_lrn$param_set$values$use_dropout = TRUE
  po_lrn$param_set$values$use_batchnorm = TRUE
  po_lrn$param_set$values$epochs = 3L
  po_lrn$param_set$values$layer_units = integer()

  pipe = po_imp %>>% po_lrn
  expect_true(!pipe$is_trained)
  pipe$train(mlr_tasks$get("pima"))
  expect_true(pipe$is_trained)
  prd = pipe$predict(mlr_tasks$get("pima"))
  expect_class(prd[[1]], "PredictionClassif")

  po_lrn = PipeOpLearner$new(lrn("classif.kerasff"))
  po_lrn$param_set$values$epochs = 3L
  po_lrn$param_set$values$layer_units = c(10, 5)
  po_lrn$param_set$values$use_embedding = FALSE
  pipe = po_imp %>>% po_lrn
  expect_true(!pipe$is_trained)
  pipe$train(mlr_tasks$get("pima"))
  expect_true(pipe$is_trained)
  prd = pipe$predict(mlr_tasks$get("pima"))
  expect_class(prd[[1]], "PredictionClassif")
  k_clear_session()
})

test_that("can fit with binary_crossentropy", {
  skip_on_os("solaris")
  skip_if_not(require("mlr3pipelines"))
  po_imp = PipeOpImputeMedian$new()
  po_lrn = PipeOpLearner$new(lrn("classif.kerasff"))
  po_lrn$param_set$values$epochs = 10L
  po_lrn$param_set$values$layer_units = c(12L, 12L)
  po_lrn$param_set$values$loss = "binary_crossentropy"
  pipe = po_imp %>>% po_lrn
  pipe$train(mlr_tasks$get("pima"))

  expect_list(pipe$state)
  expect_list(pipe$pipeops$classif.keras$state$model)
  prd = pipe$predict(mlr_tasks$get("pima"))
  expect_class(prd[[1]], "PredictionClassif")

  pipe$pipeops$classif.keras$learner$predict_type = "prob"
  prd2 = pipe$predict(mlr_tasks$get("pima"))
  expect_class(prd2[[1]], "PredictionClassif")
  expect_matrix(prd2[[1]]$prob, nrows = 768L, ncols = 2L)
  expect_true(all(prd[[1]]$response == prd2[[1]]$response))
  k_clear_session()
})