org.deeplearning4j.nn.conf.memory

Class MemoryReport

java.lang.Object

org.deeplearning4j.nn.conf.memory.MemoryReport

Direct Known Subclasses:

LayerMemoryReport, NetworkMemoryReport

public abstract class MemoryReport
extends Object

Field Summary

Fields

Modifier and Type	Field and Description
static Map<CacheMode,Long>	CACHE_MODE_ALL_ZEROS A simple Map containing all zeros for each CacheMode key

Constructor Summary

Constructors

Constructor and Description
MemoryReport()

Method Summary

Modifier and Type	Method and Description
static Map<CacheMode,Long>	cacheModeMapFor(long value) Get a map of CacheMode with all keys associated with the specified value
static MemoryReport	fromJson(String json)
static MemoryReport	fromYaml(String yaml)
protected int	getBytesPerElement(DataType dataType)
long	getMemoryBytes(MemoryType memoryType, int minibatchSize, MemoryUseMode memoryUseMode, CacheMode cacheMode) Get the memory estimate (in bytes) for the specified type of memory, using the current ND4J data type
abstract long	getMemoryBytes(MemoryType memoryType, int minibatchSize, MemoryUseMode memoryUseMode, CacheMode cacheMode, DataType dataType) Get the memory estimate (in bytes) for the specified type of memory
abstract String	getName() Name of the object that the memory report was generated for
abstract Class<?>	getReportClass()
long	getTotalMemoryBytes(int minibatchSize, @NonNull MemoryUseMode memoryUseMode, @NonNull CacheMode cacheMode) Get the total memory use in bytes for the given configuration (using the current ND4J data type)
abstract long	getTotalMemoryBytes(int minibatchSize, @NonNull MemoryUseMode memoryUseMode, @NonNull CacheMode cacheMode, @NonNull DataType dataType) Get the total memory use in bytes for the given configuration
String	toJson()
abstract String	toString()
String	toYaml()

Methods inherited from class java.lang.Object

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

Field Detail

CACHE_MODE_ALL_ZEROS

public static final Map<CacheMode,Long> CACHE_MODE_ALL_ZEROS

A simple Map containing all zeros for each CacheMode key

Constructor Detail

MemoryReport

public MemoryReport()

Method Detail

getReportClass

public abstract Class<?> getReportClass()

Returns:

Class that the memory report was generated for

getName

public abstract String getName()

Name of the object that the memory report was generated for

Returns:

Name of the object

getTotalMemoryBytes

public long getTotalMemoryBytes(int minibatchSize,
                                @NonNull
                                @NonNull MemoryUseMode memoryUseMode,
                                @NonNull
                                @NonNull CacheMode cacheMode)

Get the total memory use in bytes for the given configuration (using the current ND4J data type)

Parameters:

minibatchSize - Mini batch size to estimate the memory for

memoryUseMode - The memory use mode (training or inference)

cacheMode - The CacheMode to use

Returns:

The estimated total memory consumption in bytes

getTotalMemoryBytes

public abstract long getTotalMemoryBytes(int minibatchSize,
                                         @NonNull
                                         @NonNull MemoryUseMode memoryUseMode,
                                         @NonNull
                                         @NonNull CacheMode cacheMode,
                                         @NonNull
                                         @NonNull DataType dataType)

Get the total memory use in bytes for the given configuration

Parameters:

minibatchSize - Mini batch size to estimate the memory for

memoryUseMode - The memory use mode (training or inference)

cacheMode - The CacheMode to use

dataType - Nd4j datatype

Returns:

The estimated total memory consumption in bytes

getMemoryBytes

public long getMemoryBytes(MemoryType memoryType,
                           int minibatchSize,
                           MemoryUseMode memoryUseMode,
                           CacheMode cacheMode)

Get the memory estimate (in bytes) for the specified type of memory, using the current ND4J data type

Parameters:

memoryType - Type of memory to get the estimate for invites

minibatchSize - Mini batch size to estimate the memory for

memoryUseMode - The memory use mode (training or inference)

cacheMode - The CacheMode to use

Returns:

Estimated memory use for the given memory type

getMemoryBytes

public abstract long getMemoryBytes(MemoryType memoryType,
                                    int minibatchSize,
                                    MemoryUseMode memoryUseMode,
                                    CacheMode cacheMode,
                                    DataType dataType)

Get the memory estimate (in bytes) for the specified type of memory

Parameters:

memoryType - Type of memory to get the estimate for invites

minibatchSize - Mini batch size to estimate the memory for

memoryUseMode - The memory use mode (training or inference)

cacheMode - The CacheMode to use

dataType - Nd4j datatype

Returns:

Estimated memory use for the given memory type

toString

public abstract String toString()

Overrides:

toString in class Object

getBytesPerElement

protected int getBytesPerElement(DataType dataType)

cacheModeMapFor

public static Map<CacheMode,Long> cacheModeMapFor(long value)

Get a map of CacheMode with all keys associated with the specified value

Parameters:

value - Value for all keys

Returns:

Map

toJson

public String toJson()

toYaml

public String toYaml()

fromJson

public static MemoryReport fromJson(String json)

fromYaml

public static MemoryReport fromYaml(String yaml)