# 0.9.0

**Deeplearning4J**

* Workspaces feature added (faster training performance + less memory) [Link](https://app.gitbook.com/s/-LsGrpMiOeoMSFYK0VJQ-714541269/config/config-memory/config-workspaces.md)
* SharedTrainingMaster added for Spark network training (improved performance) [Link 1](https://deeplearning4j.oss.konduit.ai/distributed-deep-learning/intro), [Link 2](https://github.com/eclipse/deeplearning4j-examples/blob/master/dl4j-spark-examples/dl4j-spark/src/main/java/org/deeplearning4j/legacyExamples/mlp/MnistMLPDistributedExample.java)
* ParallelInference added - wrapper that server inference requests using internal batching and queues  [Link](https://github.com/eclipse/deeplearning4j-examples/blob/master/dl4j-examples/src/main/java/org/deeplearning4j/examples/inference/ParallelInferenceExample.java)
* ParallelWrapper now able to work with gradients sharing, in addition to existing parameters averaging mode [Link](https://github.com/eclipse/deeplearning4j-examples/blob/master/dl4j-cuda-specific-examples/src/main/java/org/deeplearning4j/examples/multigpu/GradientsSharingLenetMnistExample.java)
* VPTree performance significantly improved
* CacheMode network configuration option added - improved CNN and LSTM performance at the expense of additional memory use [Link](https://github.com/eclipse/deeplearning4j/blob/master/deeplearning4j/deeplearning4j-nn/src/main/java/org/deeplearning4j/nn/conf/CacheMode.java)
* LSTM layer added, with CuDNN support [Link](https://github.com/eclipse/deeplearning4j/blob/master/deeplearning4j/deeplearning4j-nn/src/main/java/org/deeplearning4j/nn/conf/layers/LSTM.java) (Note that the existing GravesLSTM implementation does not support CuDNN)
* New native model zoo with pretrained ImageNet, MNIST, and VGG-Face weights [Link](https://github.com/eclipse/deeplearning4j/tree/master/deeplearning4j/deeplearning4j-zoo/src/main/java/org/deeplearning4j/zoo)
* Convolution performance improvements, including activation caching
* Custom/user defined updaters are now supported [Link](https://github.com/eclipse/deeplearning4j/blob/master/nd4j/nd4j-backends/nd4j-api-parent/nd4j-api/src/main/java/org/nd4j/linalg/learning/config/IUpdater.java)
* Evaluation improvements
  * EvaluationBinary, ROCBinary classes added: for evaluation of binary multi-class networks (sigmoid + xent output layers) [Link](https://github.com/eclipse/deeplearning4j/blob/master/deeplearning4j/deeplearning4j-nn/src/main/java/org/deeplearning4j/eval/EvaluationBinary.java)
  * Evaluation and others now have G-Measure and Matthews Correlation Coefficient support; also macro + micro-averaging support for Evaluation class metrics [Link](https://github.com/eclipse/deeplearning4j/blob/master/deeplearning4j/deeplearning4j-nn/src/main/java/org/deeplearning4j/eval/EvaluationAveraging.java)
  * ComputationGraph and SparkComputationGraph evaluation convenience methods added (evaluateROC, etc)
  * ROC and ROCMultiClass support exact calculation (previous: thresholded calculation was used) [Link](https://github.com/eclipse/deeplearning4j/blob/master/deeplearning4j/deeplearning4j-nn/src/main/java/org/deeplearning4j/eval/ROC.java#L78-L81)
  * ROC classes now support area under precision-recall curve calculation; getting precision/recall/confusion matrix at specified thresholds (via PrecisionRecallCurve class) [Link](https://github.com/eclipse/deeplearning4j/blob/master/deeplearning4j/deeplearning4j-nn/src/main/java/org/deeplearning4j/eval/curves/PrecisionRecallCurve.java#L102-L193)
  * RegressionEvaluation, ROCBinary etc now support per-output masking (in addition to per-example/per-time-step masking)
  * EvaluationCalibration added (residual plots, reliability diagrams, histogram of probabilities) [Link 1](https://github.com/eclipse/deeplearning4j/blob/master/deeplearning4j/deeplearning4j-nn/src/main/java/org/deeplearning4j/eval/EvaluationCalibration.java) [Link 2](https://github.com/eclipse/deeplearning4j/blob/master/deeplearning4j/deeplearning4j-core/src/main/java/org/deeplearning4j/evaluation/EvaluationTools.java#L180-L188)
  * Evaluation and EvaluationBinary: now supports custom classification threshold or cost array [Link](https://github.com/eclipse/deeplearning4j/blob/master/deeplearning4j/deeplearning4j-nn/src/main/java/org/deeplearning4j/eval/Evaluation.java#L150-L170)
* Optimizations: updaters, bias calculation
* Network memory estimation functionality added. Memory requirements can be estimated from configuration without instantiating networks [Link 1](https://github.com/eclipse/deeplearning4j/blob/master/deeplearning4j/deeplearning4j-nn/src/main/java/org/deeplearning4j/nn/conf/MultiLayerConfiguration.java#L310-L317) [Link 2](https://github.com/eclipse/deeplearning4j/blob/master/deeplearning4j/deeplearning4j-nn/src/main/java/org/deeplearning4j/nn/conf/ComputationGraphConfiguration.java#L451-L458)
* New loss functions:
  * Mixture density loss function [Link](https://github.com/eclipse/deeplearning4j/blob/master/nd4j/nd4j-backends/nd4j-api-parent/nd4j-api/src/main/java/org/nd4j/linalg/lossfunctions/impl/LossMixtureDensity.java)
  * F-Measure loss function [Link](https://github.com/eclipse/deeplearning4j/blob/master/nd4j/nd4j-backends/nd4j-api-parent/nd4j-api/src/main/java/org/nd4j/linalg/lossfunctions/impl/LossFMeasure.java)

**ND4J**

* Workspaces feature added [Link](https://github.com/eclipse/deeplearning4j-examples/blob/master/nd4j-examples/src/main/java/org/nd4j/examples/Nd4jEx15_Workspaces.java)
* Native parallel sort was added
* New ops added: SELU/SELUDerivative, TAD-based comparisons, percentile/median, Reverse, Tan/TanDerivative, SinH, CosH, Entropy, ShannonEntropy, LogEntropy, AbsoluteMin/AbsoluteMax/AbsoluteSum, Atan2
* New distance functions added: CosineDistance, HammingDistance, JaccardDistance

**DataVec**

* MapFileRecordReader and MapFileSequenceRecordReader added [Link 1](https://github.com/deeplearning4j/DataVec/blob/master/datavec-hadoop/src/main/java/org/datavec/hadoop/records/reader/mapfile/MapFileRecordReader.java) [Link 2](https://github.com/deeplearning4j/DataVec/blob/master/datavec-hadoop/src/main/java/org/datavec/hadoop/records/reader/mapfile/MapFileSequenceRecordReader.java)
* Spark: Utilities to save and load `JavaRDD<List<Writable>>` and `JavaRDD<List<List<Writable>>` data to Hadoop MapFile and SequenceFile formats [Link](https://github.com/deeplearning4j/DataVec/blob/master/datavec-spark/src/main/java/org/datavec/spark/storage/SparkStorageUtils.java)
* TransformProcess and Transforms now support NDArrayWritables and NDArrayWritable columns
* Multiple new Transform classes

**Arbiter**

* Arbiter UI: [Link](https://github.com/eclipse/deeplearning4j-examples/blob/master/dl4j-examples/src/main/java/org/deeplearning4j/examples/arbiter/BasicHyperparameterOptimizationExample.java)
  * UI now uses Play framework, integrates with DL4J UI (replaces Dropwizard backend). Dependency issues/clashing versions fixed.
  * Supports DL4J StatsStorage and StatsStorageRouter mechanisms (FileStatsStorage, Remote UI via RemoveUIStatsStorageRouter)
  * General UI improvements (additional information, formatting fixes)