New Features - Added Scala Inference APIs

The new MXNet Scala Inference APIs offers an easy-to-use, and Scala Idiomatic and thread-safe high level APIs for performing predictions with deep learning models trained with MXNet.
New ImageClassifier class provides APIs for classification tasks on a Java BufferedImage using a pre-trained model you provide.
New ObjectDetector class provides APIs for object and boundary detections on a Java BufferedImage using a pre-trained model you provide.

New Features - Added module to import ONNX models into MXNet

Implemented model quantization by adopting the TensorFlow approach with calibration by borrowing the idea from Nvidia's TensorRT.
The focus of this work is on keeping quantized models (ConvNets for now) inference accuracy loss under control when compared to their corresponding FP32 models.
Please see the example on how to quantize a FP32 model with or without calibration. (#9552)

Implemented Exception Handling Support for Operators in MXNet.
Transports backend C++ exceptions to the different language front-ends and prevent crashes when exception is thrown during operator execution. (#9681)

Adds support for distributed mixed precision training with FP16. It supports storing of master copy of weights in float32 with the multi_precision mode of optimizers. (#10183)
Improved speed of float16 operations on x86 CPU by 8 times through F16C instruction set.
Added support for more operators to work with FP16 inputs. (#10125, #10078, #10169)
Added a tutorial on using mixed precision with FP16. (#10391)

Built-in profiler enhanced to support native Intel:registered: VTune:tm: Amplifier objects such as Task, Frame, Event, Counter and Marker from both C++ and Python -- which is also visible in the Chrome tracing view.
Runtime tracking of symbolic and imperative operators as well as memory and API calls.
Tracking and dumping of aggregate profiling data.
Profiler also no longer affects runtime performance when not in use.

Added Scala Inference APIs (#9678): MXNet Scala Inference API
Added module to import ONNX models into MXNet (#9963): Proposal: ImportExport module
Added support for Model Quantization with Calibration (#9552).
Added Exception Handling support for operators and iterators (#9681): Improved Exception Handling in MXNet
Added MKLDNN support for MXNet (#9677): MKLDNN integration
Added FP16 support for distributed training (#10183).
Added Profiling enhancements - VTune objects, individual operator profiling, C API profiling, Memory usage profiling (#8972)

Fixed tests - Flakiness/Bugs - (#9598, #9951, #10259, #10197, #10136, #10422). Please see: https://github.com/apache/incubator-mxnet/projects/9
Fix cudnn_conv and cudnn_deconv deadlock (#10392).
Fixed uncaught exception for bucketing module when symbol name not specified (#10094).
Fixed regression output layers (#9848).
Fixed crash with mx.nd.ones (#10014).
Fixed sample_multinomial crash when get_prob=True (#10413).
Fixed buggy type inference in correlation (#10135).
Fixed race condition for CPUSharedStorageManager->Free and launched workers at iter init stage to avoid frequent relaunch (#10096).
Fixed DLTensor Conversion for int64 (#10083).
Fixes for profiler (#9932, #10306)
Fixed ndarray assignment issues (#10022, #9981).
Fixed incorrect indices generated by device row sparse pull (#9887).
Fixed print_summary bug in visualization module (#9492).
Fixed cast storage support for same stypes (#10400).
Fixed a race condition in `io.LibSVMIter` with batch size is large (#10124).

Replaced std::swap_ranges with memcpy (#10351)
Implemented DepthwiseConv2dBackwardFilterKernel which is over 5x faster (#10098)
Implemented CPU LSTM Inference (#9977)
Added Layer Normalization in C++ (#10029)
Optimized Performance for rtc (#10018)
Improved CPU performance of ROIpooling operator by using OpenMP (#9958)
Accelerated the calculation of F1 (#9833)

Added copy flag for astype (#10347).
Added context parameter to Scala Infer API - ImageClassifier and ObjectDetector (#10252).
Added axes support for dropout in gluon (#10032).
Added default `ctx` to cpu for `gluon.Block.load_params` (#10160).
Added support for variable sequence length in gluon.RecurrentCell (#9934).
Added convenience fluent method for squeeze op (#9734).
Made array.reshape compatible with numpy (#9790).
Added axis support and gradient for L2norm (#9740)

Added support for multi-GPU training with "row_sparse" weights using "device" KVStore (#9987).
Added `Module.prepare` API for multi-GPU and multi-machine training with "row_sparse" weight (#10285)
Added 'deterministic' option for `contrib.SparseEmbedding` operator
Added `sparse.broadcast_mul` and `sparse.broadcast_div` with CSRNDArray and 1-D dense NDArray
Added sparse support for Custom Operator (#10374)
Added Sparse feature for Perl. (#9988)
Added force_deterministic option for sparse embedding (#9882).
Improved sparse SGD, sparse AdaGrad and sparse Adam optimizer speed on GPU by 30x (#9561, #10312, #10293, #10062).
Improved 'sparse.retain' performance on CPU by 2.5x (#9722)
Add `sparse.where` with condition being csr ndarray (#9481)
Added guide for implementing sparse ops (#10081).

Added constant parameter for gluon (#9893).
Added contrib.rand.zipfian (#9747).
Added Gluon PreLU, ELU, SELU, Swish activation layers for Gluon (#9662)
Added Squeeze Op (#9700).
Added multi-proposal operator (CPU version) and fixed bug in multi-proposal operator (GPU version) (#9939).
Added in Large-Batch SGD with a warmup, and a LARS startegy. (#8918)
Added Language Modelling datasets and Sampler (#9514).
Added instance norm and reflection padding to Gluon (#7938).
Added micro-averaging strategy for F1 metric (#9777).
Added Softsign Activation Function (#9851).
Added eye operator, for default storage type (#9770).
TVM bridge support to JIT NDArray Function by TVM (#9880).
Added float16 support for correlation operator and L2Normalization operator (#10125, #10078).
Added random shuffle implementation for NDArray (#10048).
Added load from buffer functions for CPP package (#10261).

Submodule:: Last updated by MXNet:: Last update in submodule