The RawData class is the base class for all raw data types. More...

Inheritance diagram for MyCaffe.layers.tft.RawData< T >:

Public Member Functions
	RawData (uint? nSeed, bool bOutputTargetHistorical)
	The constructor. More...

virtual bool	LoadData (Phase phase, string strPath, bool bShuffleData, int nBatchSize, int nHistoricalSteps, int nFutureSteps, double dfPctMaxLoad, int nDripRefreshRateInSec, uint nChunkCount, Log log, CancelEvent evtCancel)
	Loads all data values for the phase specified. More...

virtual int[,]	LoadBatch (Phase phase, int nBatchSize, BlobCollection< T > col, bool bEnableDebug=false, string strDebugPath=null)
	Loads a batch of data items into the BlobCollection. More...

virtual void	Add (LossArgs e, int[,] rgIdx)
	Adds the selected indexes along with the loss data (used by the BatchPerfSet to select worst cases). More...

virtual int	GetTotalSize ()
	Returns the total size of the data. More...

virtual int[]	GetShape (DataNpy< T >.OUTPUT_TYPE ot)
	Returns the shape of a given output type. More...

Protected Member Functions
virtual void	loadDataFunction (object obj)
	The virtual load data function override by the derived class to load the data in the background. More...

Protected Attributes
Data< T >	m_data
	Specifies the base data object used to store data blocks loaded from disk or database. More...

Random	m_random
	Specifies the random number generator used to shuffle the data. More...

int	m_nBatchSize
	Specifies the batch size. More...

bool	m_bOutputTargetHistorical
	Specifies to output the target historical data. More...

Properties
Random	Random `[get]`
	Specifies the random number generator used. More...

Detailed Description

The RawData class is the base class for all raw data types.

Template Parameters

T

Definition at line 221 of file DataTemporalLayer.cs.

Constructor & Destructor Documentation

◆ RawData()

MyCaffe.layers.tft.RawData< T >.RawData	(	uint?	nSeed,
		bool	bOutputTargetHistorical
	)

The constructor.

Parameters

nSeed	Specifies the random number generator seed.
bOutputTargetHistorical	Specifies to output the target historical data.

Definition at line 245 of file DataTemporalLayer.cs.

Member Function Documentation

◆ Add()

virtual void MyCaffe.layers.tft.RawData< T >.Add	(	LossArgs	e,
		int	rgIdx[,]
	)

virtual

Adds the selected indexes along with the loss data (used by the BatchPerfSet to select worst cases).

Parameters

e	Specifies the loss args.
rgIdx	Specifies the selected item/value indexes for the batch.

Reimplemented in MyCaffe.layers.tft.RawSqlData< T >.

Definition at line 324 of file DataTemporalLayer.cs.

◆ GetShape()

virtual int[] MyCaffe.layers.tft.RawData< T >.GetShape ( DataNpy< T >.OUTPUT_TYPE ot )

virtual

Returns the shape of a given output type.

Parameters

ot	Specifies the output type.

Returns: The shape returned can be used to reshape the Blob used to store the data on the GPU.

Reimplemented in MyCaffe.layers.tft.RawSqlData< T >.

Definition at line 342 of file DataTemporalLayer.cs.

◆ GetTotalSize()

virtual int MyCaffe.layers.tft.RawData< T >.GetTotalSize ( )

virtual

Returns the total size of the data.

Returns: The total size is returned.

Reimplemented in MyCaffe.layers.tft.RawSqlData< T >.

Definition at line 332 of file DataTemporalLayer.cs.

◆ LoadBatch()

virtual int[,] MyCaffe.layers.tft.RawData< T >.LoadBatch	(	Phase	phase,
		int	nBatchSize,
		BlobCollection< T >	col,
		bool	bEnableDebug = `false`,
		string	strDebugPath = `null`
	)

virtual

Loads a batch of data items into the BlobCollection.

Parameters

nBatchSize	Specifies the batch size.
col	Specifies the blob collection to load the batch into.
phase	Specifies the phase.
bEnableDebug	Optionally, specifies to enable debug output (default = false).
strDebugPath	Optionally, specifies the debug path where debug images are placed when 'EnableDebug' = true.

Returns: An array of the selected item and indexes is returned.

Reimplemented in MyCaffe.layers.tft.RawSqlData< T >.

Definition at line 314 of file DataTemporalLayer.cs.

◆ LoadData()

virtual bool MyCaffe.layers.tft.RawData< T >.LoadData	(	Phase	phase,
		string	strPath,
		bool	bShuffleData,
		int	nBatchSize,
		int	nHistoricalSteps,
		int	nFutureSteps,
		double	dfPctMaxLoad,
		int	nDripRefreshRateInSec,
		uint	nChunkCount,
		Log	log,
		CancelEvent	evtCancel
	)

virtual

Loads all data values for the phase specified.

Parameters

phase	Specifies the phase to load.
strPath	Specifies the base path for all data.
bShuffleData	Specifies to randomly select from the data.
nBatchSize	Specifies the batch size.
nHistoricalSteps	Specifies the number of historical steps.
nFutureSteps	Specifies the number of future steps.
dfPctMaxLoad	Specifies the percent of total items to load in background (default = 1, or 100%).
nDripRefreshRateInSec	Specifies the rate in seconds to refresh the data.
nChunkCount	Specifies the number of items to load on each cycle.
log	Specifies the output log.
evtCancel	Specifies the cancel event.

Reimplemented in MyCaffe.layers.tft.RawSqlData< T >, and MyCaffe.layers.tft.RawFileData< T >.

Definition at line 277 of file DataTemporalLayer.cs.

◆ loadDataFunction()

virtual void MyCaffe.layers.tft.RawData< T >.loadDataFunction ( object obj )

protectedvirtual

The virtual load data function override by the derived class to load the data in the background.

Parameters

obj	Specifies the user state.

Reimplemented in MyCaffe.layers.tft.RawFileData< T >.

Definition at line 301 of file DataTemporalLayer.cs.

Member Data Documentation

◆ m_bOutputTargetHistorical

bool MyCaffe.layers.tft.RawData< T >.m_bOutputTargetHistorical

protected

Specifies to output the target historical data.

Definition at line 238 of file DataTemporalLayer.cs.

◆ m_data

Data<T> MyCaffe.layers.tft.RawData< T >.m_data

protected

Specifies the base data object used to store data blocks loaded from disk or database.

Definition at line 226 of file DataTemporalLayer.cs.

◆ m_nBatchSize

int MyCaffe.layers.tft.RawData< T >.m_nBatchSize

protected

Specifies the batch size.

Definition at line 234 of file DataTemporalLayer.cs.

◆ m_random

Random MyCaffe.layers.tft.RawData< T >.m_random

protected

Specifies the random number generator used to shuffle the data.

Definition at line 230 of file DataTemporalLayer.cs.

Property Documentation

◆ Random

Random MyCaffe.layers.tft.RawData< T >.Random

get

Specifies the random number generator used.

Definition at line 258 of file DataTemporalLayer.cs.

The documentation for this class was generated from the following file:

C:/Data/Data/SS_Projects/Intelligence/GitHub/MyCaffe/MyCaffe.layers.tft/layers.tft/DataTemporalLayer.cs

Public Member Functions

Protected Member Functions

Protected Attributes

Properties

Detailed Description

Constructor & Destructor Documentation

◆ RawData()

Member Function Documentation

◆ Add()

◆ GetShape()

◆ GetTotalSize()

◆ LoadBatch()

◆ LoadData()

◆ loadDataFunction()

Member Data Documentation

◆ m_bOutputTargetHistorical

◆ m_data

◆ m_nBatchSize

◆ m_random

Property Documentation

◆ Random