cavis/arbiter
Ryan Nett 11bddb3825 SameDiff: Listener changes and training api update (#99)
* example api

Signed-off-by: Ryan Nett <rnett@skymind.io>

* Lambda based evaluation

Signed-off-by: Ryan Nett <rnett@skymind.io>

* lambda test

Signed-off-by: Ryan Nett <rnett@skymind.io>

* fixes

Signed-off-by: Ryan Nett <rnett@skymind.io>

* partial fixes, use get-variable listener framework, example EvaluationListener

Signed-off-by: Ryan Nett <rnett@skymind.io>

* javadoc fix and newInstance implementations

Signed-off-by: Ryan Nett <rnett@skymind.io>

* fit and evaluate methods with validation data (for fit) and listeners

Signed-off-by: Ryan Nett <rnett@skymind.io>

* output method overloads + listener args

Signed-off-by: Ryan Nett <rnett@skymind.io>

* history and evaluation helpers

Signed-off-by: Ryan Nett <rnett@skymind.io>

* fixes

Signed-off-by: Ryan Nett <rnett@skymind.io>

* more fixes

Signed-off-by: Ryan Nett <rnett@skymind.io>

* FitConfig and added getters and setters

Signed-off-by: Ryan Nett <rnett@skymind.io>

* javadocs

Signed-off-by: Ryan Nett <rnett@skymind.io>

* fixes, javadoc, added activations to history, added latest activation listener

Signed-off-by: Ryan Nett <rnett@skymind.io>

* fixes, start of tests

Signed-off-by: Ryan Nett <rnett@skymind.io>

* fixes and updates

Signed-off-by: Ryan Nett <rnett@skymind.io>

* newInstance fixes, tests

Signed-off-by: Ryan Nett <rnett@skymind.io>

* test fixes

Signed-off-by: Ryan Nett <rnett@skymind.io>

* javadocs, getters with SDVariable overrides, CustomEvaluation fix

Signed-off-by: Ryan Nett <rnett@skymind.io>

* more operation config classes (evaluation, output, exec/single batch output), fix custom eval tests

Signed-off-by: Ryan Nett <rnett@skymind.io>

* merge fixes

Signed-off-by: Ryan Nett <rnett@skymind.io>

* fix

Signed-off-by: Ryan Nett <rnett@skymind.io>

* fixes, most old fit/evaluate/output methods use the builders

Signed-off-by: Ryan Nett <rnett@skymind.io>

* fixes

Signed-off-by: Ryan Nett <rnett@skymind.io>

* numerous fixes/cleanup

Signed-off-by: Ryan Nett <rnett@skymind.io>

* fixes

Signed-off-by: Ryan Nett <rnett@skymind.io>

* javadoc

Signed-off-by: Ryan Nett <rnett@skymind.io>

* Polish round 1

Signed-off-by: AlexDBlack <blacka101@gmail.com>

* Round 2

Signed-off-by: AlexDBlack <blacka101@gmail.com>

* Formatting + round 3

Signed-off-by: AlexDBlack <blacka101@gmail.com>

* Round 4

Signed-off-by: AlexDBlack <blacka101@gmail.com>
2019-08-10 15:30:31 +10:00
..
.github Eclipse Migration Initial Commit 2019-06-06 15:21:15 +03:00
arbiter-core Merge master to upstream (#7945) 2019-06-27 18:37:04 +03:00
arbiter-deeplearning4j SameDiff: Listener changes and training api update (#99) 2019-08-10 15:30:31 +10:00
arbiter-server Eclipse Migration Initial Commit 2019-06-06 15:21:15 +03:00
arbiter-ui Eclipse Migration Initial Commit 2019-06-06 15:21:15 +03:00
ci Eclipse Migration Initial Commit 2019-06-06 15:21:15 +03:00
contrib Eclipse Migration Initial Commit 2019-06-06 15:21:15 +03:00
.travis.yml Eclipse Migration Initial Commit 2019-06-06 15:21:15 +03:00
README.md Eclipse Migration Initial Commit 2019-06-06 15:21:15 +03:00
buildmultiplescalaversions.sh Eclipse Migration Initial Commit 2019-06-06 15:21:15 +03:00
pom.xml Eclipse Migration Initial Commit 2019-06-06 15:21:15 +03:00

README.md

Arbiter

A tool dedicated to tuning (hyperparameter optimization) of machine learning models. Part of the DL4J Suite of Machine Learning / Deep Learning tools for the enterprise.

Modules

Arbiter contains the following modules:

  • arbiter-core: Defines the API and core functionality, and also contains functionality for the Arbiter UI
  • arbiter-deeplearning4j: For hyperparameter optimization of DL4J models (MultiLayerNetwork and ComputationGraph networks)

Hyperparameter Optimization Functionality

The open-source version of Arbiter currently defines two methods of hyperparameter optimization:

  • Grid search
  • Random search

For optimization of complex models such as neural networks (those with more than a few hyperparameters), random search is superior to grid search, though Bayesian hyperparameter optimization schemes For a comparison of random and grid search methods, see Random Search for Hyper-parameter Optimization (Bergstra and Bengio, 2012).

Core Concepts and Classes in Arbiter for Hyperparameter Optimization

In order to conduct hyperparameter optimization in Arbiter, it is necessary for the user to understand and define the following:

  • Parameter Space: A ParameterSpace<P> specifies the type and allowable values of hyperparameters for a model configuration of type P. For example, P could be a MultiLayerConfiguration for DL4J
  • Candidate Generator: A CandidateGenerator<C> is used to generate candidate models configurations of some type C. The following implementations are defined in arbiter-core:
    • RandomSearchCandidateGenerator
    • GridSearchCandidateGenerator
  • Score Function: A ScoreFunction<M,D> is used to score a model of type M given data of type D. For example, in DL4J a score function might be used to calculate the classification accuracy from a DataSetIterator
    • A key concept here is that they score is a single numerical (double precision) value that we either want to minimize or maximize - this is the goal of hyperparameter optimization
  • Termination Conditions: One or more TerminationCondition instances must be provided to the OptimizationConfiguration. TerminationCondition instances are used to control when hyperparameter optimization should be stopped. Some built-in termination conditions:
    • MaxCandidatesCondition: Terminate if more than the specified number of candidate hyperparameter configurations have been executed
    • MaxTimeCondition: Terminate after a specified amount of time has elapsed since starting the optimization
  • Result Saver: The ResultSaver<C,M,A> interface is used to specify how the results of each hyperparameter optimization run should be saved. For example, whether saving should be done to local disk, to a database, to HDFS, or simply stored in memory.
    • Note that ResultSaver.saveModel method returns a ResultReference object, which provides a mechanism for re-loading both the model and score from wherever it may be saved.
  • Optimization Configuration: An OptimizationConfiguration<C,M,D,A> ties together the above configuration options in a fluent (builder) pattern.
  • Candidate Executor: The CandidateExecutor<C,M,D,A> interface provides a layer of abstraction between the configuration and execution of each instance of learning. Currently, the only option is the LocalCandidateExecutor, which is used to execute learning on a single machine (in the current JVM). In principle, other execution methods (for example, on Spark or cloud computing machines) could be implemented.
  • Optimization Runner: The OptimizationRunner uses an OptimizationConfiguration and a CandidateExecutor to actually run the optimization, and save the results.

Optimization of DeepLearning4J Models

(This section: forthcoming)