ai.djl.mxnet.engine

Class MxNDManager

java.lang.Object

ai.djl.ndarray.BaseNDManager

ai.djl.mxnet.engine.MxNDManager

All Implemented Interfaces:

ai.djl.ndarray.NDManager, java.lang.AutoCloseable

public class MxNDManager
extends ai.djl.ndarray.BaseNDManager

MxNDManager is the MXNet implementation of NDManager.

Field Summary

Fields inherited from class ai.djl.ndarray.BaseNDManager

closed, device, parent, resources, uid

Method Summary

Modifier and Type	Method and Description
java.nio.ByteBuffer	allocateDirect(int capacity)
ai.djl.ndarray.NDArray	arange(float start, float stop, float step, ai.djl.ndarray.types.DataType dataType, ai.djl.Device dev)
MxNDArray	create(com.sun.jna.Pointer handle) Creates an MxNDArray with the given Native Memory Pointer and attaches to this manager.
MxSparseNDArray	create(com.sun.jna.Pointer handle, ai.djl.ndarray.types.SparseFormat fmt) Creates a sparse MxNDArray with the given Native Memory Pointer and attaches to this manager.
MxNDArray	create(ai.djl.ndarray.types.Shape shape, ai.djl.ndarray.types.DataType dataType, ai.djl.Device dev)
MxSparseNDArray	createCSR(java.nio.Buffer data, long[] indptr, long[] indices, ai.djl.ndarray.types.Shape shape, ai.djl.Device dev)
MxSparseNDArray	createRowSparse(java.nio.Buffer data, ai.djl.ndarray.types.Shape dataShape, long[] indices, ai.djl.ndarray.types.Shape shape, ai.djl.Device dev)
ai.djl.ndarray.NDArray	eye(int rows, int cols, int k, ai.djl.ndarray.types.DataType dataType, ai.djl.Device dev)
ai.djl.engine.Engine	getEngine()
void	invoke(java.lang.String operation, ai.djl.ndarray.NDArray[] src, ai.djl.ndarray.NDArray[] dest, ai.djl.util.PairList<java.lang.String,?> params)
ai.djl.ndarray.NDArray	invoke(java.lang.String operation, ai.djl.ndarray.NDArray[] src, ai.djl.util.PairList<java.lang.String,?> params) An engine specific generic invocation to native operator.
ai.djl.ndarray.NDArray	invoke(java.lang.String operation, ai.djl.ndarray.NDArray src, ai.djl.util.PairList<java.lang.String,?> params) An engine specific generic invocation to native operator.
void	invoke(java.lang.String operation, ai.djl.ndarray.NDList src, ai.djl.ndarray.NDList dest, ai.djl.util.PairList<java.lang.String,?> params) An engine specific generic invocation to native operator.
ai.djl.ndarray.NDList	invoke(java.lang.String operation, ai.djl.ndarray.NDList src, ai.djl.util.PairList<java.lang.String,?> params)
ai.djl.ndarray.NDArray	invoke(java.lang.String operation, ai.djl.util.PairList<java.lang.String,?> params) An engine specific generic invocation to native operator.
ai.djl.ndarray.NDArray	linspace(float start, float stop, int num, boolean endpoint, ai.djl.Device dev)
ai.djl.ndarray.NDList	load(java.nio.file.Path path, ai.djl.Device device)
MxNDManager	newSubManager()
MxNDManager	newSubManager(ai.djl.Device dev)
ai.djl.ndarray.NDArray	ones(ai.djl.ndarray.types.Shape shape, ai.djl.ndarray.types.DataType dataType, ai.djl.Device dev)
ai.djl.ndarray.NDArray	randomMultinomial(int n, ai.djl.ndarray.NDArray pValues)
ai.djl.ndarray.NDArray	randomMultinomial(int n, ai.djl.ndarray.NDArray pValues, ai.djl.ndarray.types.Shape shape)
ai.djl.ndarray.NDArray	randomNormal(float loc, float scale, ai.djl.ndarray.types.Shape shape, ai.djl.ndarray.types.DataType dataType, ai.djl.Device dev)
ai.djl.ndarray.NDArray	randomUniform(float low, float high, ai.djl.ndarray.types.Shape shape, ai.djl.ndarray.types.DataType dataType, ai.djl.Device dev)
ai.djl.ndarray.NDArray	zeros(ai.djl.ndarray.types.Shape shape, ai.djl.ndarray.types.DataType dataType, ai.djl.Device dev)

Methods inherited from class ai.djl.ndarray.BaseNDManager

attach, close, debugDump, detach, getDevice, getParentManager, isOpen, toString

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, wait, wait, wait

Methods inherited from interface ai.djl.ndarray.NDManager

arange, arange, arange, arange, arange, arange, arange, create, create, create, create, create, create, create, create, create, create, create, create, create, create, create, create, create, create, create, create, create, create, create, create, create, create, create, create, create, createCSR, createRowSparse, decode, decode, eye, eye, eye, linspace, linspace, linspace, load, newBaseManager, newBaseManager, newBaseManager, ones, ones, randomNormal, randomNormal, randomUniform, zeros, zeros

Method Detail

allocateDirect

public java.nio.ByteBuffer allocateDirect(int capacity)

create

public MxNDArray create(com.sun.jna.Pointer handle)

Creates an MxNDArray with the given Native Memory Pointer and attaches to this manager.

Parameters:

handle - the array's native memory pointer

Returns:

the created array

create

public MxSparseNDArray create(com.sun.jna.Pointer handle,
                              ai.djl.ndarray.types.SparseFormat fmt)

Creates a sparse MxNDArray with the given Native Memory Pointer and attaches to this manager.

Parameters:

handle - the array's native memory pointer

fmt - the sparse format to use

Returns:

the created array

create

public MxNDArray create(ai.djl.ndarray.types.Shape shape,
                        ai.djl.ndarray.types.DataType dataType,
                        ai.djl.Device dev)

createCSR

public MxSparseNDArray createCSR(java.nio.Buffer data,
                                 long[] indptr,
                                 long[] indices,
                                 ai.djl.ndarray.types.Shape shape,
                                 ai.djl.Device dev)

createRowSparse

public MxSparseNDArray createRowSparse(java.nio.Buffer data,
                                       ai.djl.ndarray.types.Shape dataShape,
                                       long[] indices,
                                       ai.djl.ndarray.types.Shape shape,
                                       ai.djl.Device dev)

load

public ai.djl.ndarray.NDList load(java.nio.file.Path path,
                                  ai.djl.Device device)

zeros

public ai.djl.ndarray.NDArray zeros(ai.djl.ndarray.types.Shape shape,
                                    ai.djl.ndarray.types.DataType dataType,
                                    ai.djl.Device dev)

ones

public ai.djl.ndarray.NDArray ones(ai.djl.ndarray.types.Shape shape,
                                   ai.djl.ndarray.types.DataType dataType,
                                   ai.djl.Device dev)

arange

public ai.djl.ndarray.NDArray arange(float start,
                                     float stop,
                                     float step,
                                     ai.djl.ndarray.types.DataType dataType,
                                     ai.djl.Device dev)

eye

public ai.djl.ndarray.NDArray eye(int rows,
                                  int cols,
                                  int k,
                                  ai.djl.ndarray.types.DataType dataType,
                                  ai.djl.Device dev)

linspace

public ai.djl.ndarray.NDArray linspace(float start,
                                       float stop,
                                       int num,
                                       boolean endpoint,
                                       ai.djl.Device dev)

randomUniform

public ai.djl.ndarray.NDArray randomUniform(float low,
                                            float high,
                                            ai.djl.ndarray.types.Shape shape,
                                            ai.djl.ndarray.types.DataType dataType,
                                            ai.djl.Device dev)

randomNormal

public ai.djl.ndarray.NDArray randomNormal(float loc,
                                           float scale,
                                           ai.djl.ndarray.types.Shape shape,
                                           ai.djl.ndarray.types.DataType dataType,
                                           ai.djl.Device dev)

randomMultinomial

public ai.djl.ndarray.NDArray randomMultinomial(int n,
                                                ai.djl.ndarray.NDArray pValues,
                                                ai.djl.ndarray.types.Shape shape)

randomMultinomial

public ai.djl.ndarray.NDArray randomMultinomial(int n,
                                                ai.djl.ndarray.NDArray pValues)

newSubManager

public MxNDManager newSubManager()

newSubManager

public MxNDManager newSubManager(ai.djl.Device dev)

invoke

public void invoke(java.lang.String operation,
                   ai.djl.ndarray.NDArray[] src,
                   ai.djl.ndarray.NDArray[] dest,
                   ai.djl.util.PairList<java.lang.String,?> params)

invoke

public ai.djl.ndarray.NDList invoke(java.lang.String operation,
                                    ai.djl.ndarray.NDList src,
                                    ai.djl.util.PairList<java.lang.String,?> params)

invoke

public void invoke(java.lang.String operation,
                   ai.djl.ndarray.NDList src,
                   ai.djl.ndarray.NDList dest,
                   ai.djl.util.PairList<java.lang.String,?> params)

An engine specific generic invocation to native operator.

You should avoid using this function if possible. Since this function is engine specific, using this API may cause portability issues. A native operation may not be compatible between each version.

Parameters:

operation - the native operation to perform

src - the NDList of source NDArray

dest - the NDList to save output to

params - the parameters to be passed to the native operator

Throws:

java.lang.IllegalArgumentException - if operation is not supported by Engine

ai.djl.engine.EngineException - if operation failed in native engine

invoke

public ai.djl.ndarray.NDArray invoke(java.lang.String operation,
                                     ai.djl.ndarray.NDArray[] src,
                                     ai.djl.util.PairList<java.lang.String,?> params)

An engine specific generic invocation to native operator.

You should avoid using this function if possible. Since this function is engine specific, using this API may cause portability issues. A native operation may not be compatible between each version.

Parameters:

operation - the native operation to perform

src - the array of source NDArray

params - the parameters to be passed to the native operator

Returns:

the output array of NDArray

Throws:

java.lang.IllegalArgumentException - if operation is not supported by Engine

ai.djl.engine.EngineException - if operation failed in native engine

invoke

public ai.djl.ndarray.NDArray invoke(java.lang.String operation,
                                     ai.djl.ndarray.NDArray src,
                                     ai.djl.util.PairList<java.lang.String,?> params)

An engine specific generic invocation to native operator.

You should avoid using this function if possible. Since this function is engine specific, using this API may cause portability issues. A native operation may not be compatible between each version.

Parameters:

operation - the native operation to perform

src - the source NDArray

params - the parameters to be passed to the native operator

Returns:

the output array of NDArray

Throws:

java.lang.IllegalArgumentException - if operation is not supported by Engine

ai.djl.engine.EngineException - if operation failed in native engine

invoke

public ai.djl.ndarray.NDArray invoke(java.lang.String operation,
                                     ai.djl.util.PairList<java.lang.String,?> params)

An engine specific generic invocation to native operator.

You should avoid using this function if possible. Since this function is engine specific, using this API may cause portability issues. A native operation may not be compatible between each version.

Parameters:

operation - the native operation to perform

params - the parameters to be passed to the native operator

Returns:

the output array of NDArray

Throws:

java.lang.IllegalArgumentException - if operation is not supported by Engine

ai.djl.engine.EngineException - if operation failed in native engine

getEngine

public ai.djl.engine.Engine getEngine()