Open
Description
Description
Hi, I was confused with when using "classif.kerascnn" method the "Feature Types" is "-". And the error is not supporting numeric feature type when training the learner.
The following is a demonstration：
Reproducible example
> lrn = lrn("classif.kerascnn")
> lrn
<LearnerClassifKeras:classif.keras>
* Model: -
* Parameters: epochs=100, callbacks=<list>, validation_split=0.3333,
batch_size=128, low_memory=FALSE, verbose=0, application=<function>,
optimizer=<keras.optimizer_v2.adam.Adam>, loss=categorical_crossentropy,
metrics=accuracy, cl_layer_units=1024, output_activation=softmax,
unfreeze_n_last_layers=5, validation_fraction=0.2
* Packages: mlr3, keras
* Predict Types: [response], prob
* Feature Types: -
* Properties: multiclass, twoclass
>lrn$train(tsk("iris"))
Error: <TaskClassif:iris> has the following unsupported feature types: numeric
Metadata
Metadata
Assignees
Labels
No labels