Soft-labeling means to annotate data with labels that were predicited by a machine classifier.
As they were not evaluated by a human, you might call them "soft".

Within nkululeko, this can be achieved in two ways:
a) save the test/evaluation set as a new database
or
b) use the predict module to predict labels for the test set.

a) save the test set

You simply specify a name in the EXP section to save the test predictions like this:

[EXP]
...
save_test = ./my_test_predictions.csv

This will store the new database to a file called my_test_predictions.csv in the folder where the python was called.

b) use the predict module

After each nkululeko call, you will find the test set in CSV file format in the store folder inside the project folder, named testdf.csv.

Use this file as input to the predict module to predict labels.