The Div macro wraps content in a div tag with optional class and styles. This allows us to use macros such as the Style macro below.
There are 2 phases to applying Deep Learning to a ML problem, the first phase is where a neural network is created and trained using training data to generate pre-trained model and In the second phase, this pre-trained model is put to work by running inference(forward-pass) on new data in the customer’s application in production. Model Creation and Training is typically performed by Data Scientists who prefer using Python as a primary language which provides rich set of libraries(numpy, pandas, pillow) etc., to setup the training pipeline. MXNet already has very good support for Python to quickly prototype and develop models.
Inference on the other hand is run and managed by Software Engineers in a production eco-system which is built with tools and frameworks that use Java/Scala as a primary Language.
Inference on a trained model has two different use-cases:
Batch Inference is performed on big data platforms such as Spark using Scala or Java while Real time Inference is typically performed and deployed on popular web frameworks such as Tomcat, Netty, Jetty, etc. which use Java.
With this project, we want to build a new set of APIs which are Java friendly, compatible with Java 7+, are easy to use for inference, and lowers the entry barrier of consuming MXNet for production use-cases.
As a user, I’d like to have a Java Inference API that allows me to use deep learning models from my existing Java application.
As a user, I'd like for the new Java Inference API to be thread safe.
As a user, I’d like for the new Java inference API to be idiomatic and easy to use so that I can quickly learn to deploy models.
As a user, I’d like for the new Java Inference API to introduce as few dependencies as possible so that it’s easy to add into my existing environment.
As a user, I’d like for the new Java API to have full test coverage so that I can be confident in it’s stability.
As a user, I’d like for the Java API to be as performant as possible with the performance results measured and available so that I can compare implementations and make informed decisions.
As a user, I would like for the new Java Inference API to support RNNs so that the models I’ve trained can be deployed to our production environments.
As a user already familiar with MXNet, I’d like for the new API to be similar to existing implementations so that it’s easy for me to use.
As a user, I'd like to have examples and tutorials available to help learn how to use the new Java Inference API.
The proposed implementation for the new Java API is to create a Java friendly wrapper around the existing Scala API. The Scala API is already fully implemented and is undergoing significant improvements (most notably simplifying the memory management of off-heap memory). By utilizing the existing Scala API, the development effort require for the new Java API is greatly decreased. Additionally, the Java API would automatically (or with minimal work) benefit from new features and code improvements allowing for development efforts to remain focused. This is a very similar approach to how Apache Spark developed their Java API.
Since both Java and Scala are JVM languages, it is already possible for the Scala bindings to be called from Java code by loading the jar into the classpath. Due to differences in the languages, this process is currently very painful for users to implement. Most notably, the difficulty comes from the liberal use of default values in the Scala code being unsupported by Java and converting between Java/Scala collections. To improve upon this experience, a Java wrapper would be created which will call the Scala bindings. The wrapper would be designed so that it abstracts away the complexities of the Java/Scala interaction by automating the conversions, simplifying the method calls, and making the API more idiomatic for the Java inferencing use case.
Milestone 1: Initial release with support for all existing Scala Inference APIs. Includes integration into the existing CI, working examples, tutorials, documentation, benchmarking, and integrations into Maven distribution pipeline.
Milestone 2: General improvements to Inference API, improved better support specific use cases, and add sparse support (required for RNNs).
Milestone 3+: ?? (Ideas include: auto grad, exposing module api, control flow support)
Converting Java collections into Scala collections - Scala and Java use different collections. Generally, these can be converted through the scala.collection.JavaConverters library. Ideally, this will be done automatically on behalf of the user. The Java methods should take Java collections, do the necessary conversion, then call the corresponding Scala method.
Java doesn’t support methods with default arguments - The current Scala implementation makes liberal use of default arguments. For class instantiation, a simple builder pattern will work. Class methods with default values will likely need to be overloaded.
Limited by existing Scala Inference API - The current Scala Inference API is lacking support for some models such as RNNs. Since this API will be utilized by the new Java Inference API, it will be necessary to improve and expand the Scala Inference API. This work can be done in parallel and should undergo it’s own design process. On the plus side this will serve as a forcing function to improve the Scala API.
Performance should be very similar to Scala. Since both are JVM languages doing inference will be calling the same byte code from Java as it is in Scala. The only known issue which will cause a performance difference is converting the Java collections into Scala collections. Preliminary testing with simple models shows negligible to nonexistent impact to performance. Java performance should be measured via Benchmark Scripts in a manner similar to how it's measured in Scala. More details on Scala benchmarks are available here.
Preliminary comparison results are a WIP and will be added soon.
The new Java inference API can be distributed alongside the existing Scala API. Currently, the Scala API is distributed via a jar file using a Maven repository. There is ongoing work to automate this process and ideally this work will include the new Java API as well. The design for the Automated Scala Release is available here. Releases for the Java Inference API will be aligned with the MXNet release schedule and follow the same versioning.
The existing Scala Inference API will need to be expanded and improved. These changes will need to undergo their own design process and can easily be incorporated into the new Java API. Although these improvements are not a requirement to begin working on the the Java API, ideally it will be done in parallel so that the Java API will be more useful upon release.
Known improvements to could made to the Scala API include:
The Java Inference API will be a wrapper around the high level Scala Inference interface. Here is an example of what the Java wrapper will look like for the Scala inference Predictor class.
/**
* Implementation of prediction routines.
*
* @param modelPathPrefix Path prefix from where to load the model artifacts.
* These include the symbol, parameters, and synset.txt
* Example: file://model-dir/resnet-152 (containing
* resnet-152-symbol.json, resnet-152-0000.params, and synset.txt).
* @param inputDescriptors Descriptors defining the input node names, shape,
* layout and type parameters
* <p>Note: If the input Descriptors is missing batchSize
* ('N' in layout), a batchSize of 1 is assumed for the model.
* @param contexts Device contexts on which you want to run inference; defaults to CPU
* @param epoch Model epoch to load; defaults to 0
*/
Predictor(String modelPathPrefix, List<DataDesc> inputDescriptors,
List<Context> Contexts, int epoch)
/**
* Predict using NDArray as input
* This method is useful when the input is a batch of data
* Note: User is responsible for managing allocation/deallocation of input/output NDArrays.
*
* @param inputBatch List of NDArrays
* @return Output of predictions as NDArrays
*/
List <NDArray> predictWithNDArray(List <NDArray> inputBatch) |
A primary goal of the Java Inference API is to provide a simple means for Java users to load and do inference on an existing model. Ideally, this will typically be as simple as defining the context (cpu vs gpu) to be used, defining what the input will look like, and setting up the model that will be used. After setting up the model like this, it should be simple to do input on the model.
/*
* Psudeocode for how ObjectDetector Class can be used to do SSD detection
* A full working SSD example will be included in the release.
*/
// Set the context to be used
List<Context> context = new ArrayList<Context>();
context.add(Context.cpu());
// Define the shape and data type of the input
Shape inputShape = new Shape(new int[] {1, 3, 512, 512});
List<DataDesc> inputDescriptors = new ArrayList<DataDesc>();
inputDescriptors.add(new DataDesc("data", inputShape, DType.Float32(), "NCHW"));
// Instantiate the object detector with the model, input descriptors, context, and epoch
JavaObjectDetector objDetector = new JavaObjectDetector(modelPathPrefix, inputDescriptors, context, 0);
// Load an image and run inference on it
BufferedImage img = JavaImageClassifier.loadImageFromFile(inputImagePath);
objDetector.imageObjectDetect(img, 3); |
On Java side:
On Scala side:
Construct interfaces in Java that define all types that will be passed between Java and Scala. Place these interfaces into a project that can be shared between the Java portions of code and the Scala portions of code. By limiting the features used in the integration points, there won’t be any feature mismatch issues. (referred from "scala-in-depth" page 242)
Writing a Java Inference API that directly calls the native code - Doing this would be designing and implementing a Java Inference API that will interact with the native code using the existing JNI code. The API would be designed to make Java inferencing simple and idiomatic. The existing JNI code could be shared by both the existing Scala API and the new Java Inference API. The biggest drawback to this approach is that it involves a significant amount of duplicate work that would be very difficult to maintain with current resources.
2. Disadvantages
Adopt Java as the primary JVM language - This approach is basically to spend a very significant effort rewriting the entire Scala API into Java. After that was done we could begin adding support for other JVM languages using Java as a base and eventually the current Scala API could be deprecated. Obviously this involves a very significant upfront effort. Long-term it would be reasonable to expect improved performance across all JVM languages (since benchmarks generally show Java to outperform Scala) and it would likely be easier to add support for other JVM languages. The performance gains would likely be offset tremendously by the fact that most of the workload is done in the C++ native code and not in the JVM.
1. Advantages
Likely to see better performance across all JVM languages.
Easier to add support for more JVM languages in the future.
2. Disadvantages
Tremendous amount of upfront effort.
Scala API already exists and has been well received.