Execution format

catboost fit -f <file path> [--cv <value>] [--cv-rand <value>] [other parameters]

Options
...
--cv 
Perform cross validation to score the model. Supports two modes. Classical: perform cross validation to score the model by excluding a small fold from the training dataset and using it for testing. Inverted: perform cross validation to score the model by excluding a small fold from the validation dataset and using it for training.

Format for specifying values:

<cvtype>:<n>;<k>
cvtype is either Classical or Inverted.
 n is the identifier of the fold to exclude.
 k is the number of folds to split the input data into.
The inequality n < k must be true.

For example, to exclude the fold indexed 13 from the training dataset and use it for testing, set the value Classical:13;15. To exclude the fold indexed 13 from the validation dataset and use it for training, set the value Inverted:13;15.

The data is randomly shuffled before splitting.

--cv-rand
...
It must be used with --cv <value> parameter.