Infer.NET user guide : Learners : Bayes Point Machine classifiers : Command-line runners

Incremental Training

A Bayes Point Machine is trained incrementally using the TrainIncremental module, both in binary and multi-class classification. The TrainIncremental module reads a serialized trained classifier and a training set and returns an incrementally trained classifier, which can then be used to make predictions or train once more (incrementally).

The TrainIncremental module has the following command-line arguments:

Required arguments

• training-set: The file with training data containing ground truth labels and features in the format described earlier.
• input-model: The file from which a previously trained Bayes Point Machine classifier will be loaded.
• model: The file to which the trained Bayes Point Machine classifier will be saved.

Optional arguments

• iterations: The number of training algorithm iterations (defaults to 30).
• batches: The number of batches into which the training data is split (defaults to 1).

For more information about the command-line arguments, see Settings. A more detailed explanation of incremental training is available here.

Example

Learner Classifier BinaryBayesPointMachine TrainIncremental   
    --input-model trained-binary-bpm.bin --training-set training-set.dat   
    --model incrementally-trained-binary-bpm.bin --iterations 15 --batches 3   

Learner Classifier MulticlassBayesPointMachine TrainIncremental   
    --input-model trained-multiclass-bpm.bin --training-set training-set.dat   
    --model incrementally-trained-multiclass-bpm.bin --iterations 15 --batches 3