MXNet. Deep Learning API.

지속가능티끌/Data.Math.Phys

i.got.it 2016. 11. 6. 13:06

Overview

주요특징.

- Flexible and Efficient Library for Deep Learning

- mxnet 소스 : C++11

- Portable 실행 : Runs on CPUs or GPUs, on clusters, servers, desktops, or mobile phones

- Cloud 실행 : Supports distributed training on multiple CPU/GPU machines, including AWS, GCE, Azure, and Yarn clusters

- API 언어 : C++ , R , Python, Javascript, Scala, Julia, Matlab.

mxnet main site : http://mxnet.io/

pdf 문서 :

MXNet Function Block

Above image shows major modules/components of the MXNet system and their interaction. The modules are:

Runtime Dependency Engine: Schedules and executes the operations according to their read/write dependency.
Storage Allocator: Efficiently allocate and recycles memory blocks for GPU and CPU.
Resource Manager: Manage global resources such as random number generator, temporal space.
NDArray: Dynamic asynchronous n-dimensional arrays, provide flexible imperative programs for MXNet.
Symbolic Execution: Static symbolic graph executor, provide efficient symbolic graph execution and optimization.
Operator: Operators that defines static forward and gradient calculation(backprop).
SimpleOp: Operators that extend to NDArray operators and symbolic operators in a unified fashion.
Symbol Construction: Symbolic construction, provide a way to construct computation graph(net configuration)
KVStore: Key-value store interface for easy parameter synchronizations.
Data Loading(IO): Efficient distributed data loading and augmentation

///1057.