Describe the use case
Training multiple models with shuffled training data (k-folds) can reveals information about our data sets. For example, one fold may be much less accurate than other folds telling us that we might need to increase the total number of samples.

Describe the solution you'd like
For the CLI, it would be useful if we could have a parameter to train with k-fold cross validation.

Describe alternatives you've considered
k-fold can be achieved with the python API but not as convenient as the CLI.

Additional context
I can provide the sample code I've been using for the API.