TrainingSession: New Training Configuration

In Microsoft/CNTK-R: R Interface to the Microsoft Cognitive Toolkit (CNTK)

Description

The instance of the class should be created by using training_session() function. A training session trains a model using the specified trainer and configs. Different aspects of training such as data sources, checkpointing, cross validation, progress printing can be configured using the corresponding config classes.

Usage

TrainingSession(trainer, mb_source, mb_size, model_inputs_to_streams,
  max_samples, progress_frequency, checkpoint_config, cv_config, test_config)

Arguments

trainer	(Trainer): trainer
mb_source	(MinibatchSource): minibatch source used for training
mb_size	(minibatch_size_schedule or int): minibatch size schedule for training
model_inputs_to_streams	(dict): mapping between input variables and input streams
max_samples	(int): maximum number of samples used for training
progress_frequency	(int): frequency in samples for aggregated progress printing
checkpoint_config	(CheckpointConfig): checkpoint configuration
cv_config	(CrossValidationConfig): cross validation configuration
test_config	(TestConfig): test configuration

Details

****** Associated Functions: ******

on_train_cross_validation_end

train_on_session

Microsoft/CNTK-R documentation built on May 28, 2019, 1:52 p.m.