R/feature_spec.R

dataset_use_spec

Transform the dataset using the provided spec.

Description

Prepares the dataset to be used directly in a model.The transformed dataset is prepared to return tuples (x,y) that can be used directly in Keras.

Usage

 
dataset_use_spec(dataset, spec) 

Arguments

Arguments Description
dataset A TensorFlow dataset.
spec A feature specification created with feature_spec().

Value

A TensorFlow dataset.

Examples

library(tfdatasets) 
data(hearts) 
hearts <- tensor_slices_dataset(hearts) %>% dataset_batch(32) 
 
# use the formula interface 
spec <- feature_spec(hearts, target ~ age) %>% 
  step_numeric_column(age) 
 
spec_fit <- fit(spec) 
final_dataset <- hearts %>% dataset_use_spec(spec_fit) 

See Also

  • feature_spec() to initialize the feature specification.

  • fit.FeatureSpec() to create a tensorflow dataset prepared to modeling.

  • steps to a list of all implemented steps.

    Other Feature Spec Functions: feature_spec(), fit.FeatureSpec(), step_bucketized_column(), step_categorical_column_with_hash_bucket(), step_categorical_column_with_identity(), step_categorical_column_with_vocabulary_file(), step_categorical_column_with_vocabulary_list(), step_crossed_column(), step_embedding_column(), step_indicator_column(), step_numeric_column(), step_remove_column(), step_shared_embeddings_column(), steps