MyCaffe.layers.hdf5.HDF5< T > Class Template Reference

The HDF5 object provides HDF5 dataset support to the HDF5DataLayer. More...

Inheritance diagram for MyCaffe.layers.hdf5.HDF5< T >:

Public Member Functions
	HDF5 (CudaDnn< T > cuda, Log log, string strFile)
	The constructor. More...
	HDF5 ()
	The constructor used when loading weights into a Net. More...
void	load_nd_dataset (Blob< T > blob, string strDatasetName, bool bReshape=false, int nMinDim=1, int nMaxDim=int.MaxValue, H5GroupId id=null, bool bAllowSingleItems=false)
	Creates a new dataset from an HDF5 data file. More...
void	CopyTrainedLayersFromHDF5 (Net< T > net, Log log, string strFile)
	Copy the weights from an HDF5 file into a Net. More...
void	Dispose ()
	Release all resources uses. More...

Detailed Description

The HDF5 object provides HDF5 dataset support to the HDF5DataLayer.

Template Parameters

T	Specifies the base type.

Definition at line 32 of file HDF5.cs.

Constructor & Destructor Documentation

HDF5() [1/2]

MyCaffe.layers.hdf5.HDF5< T >.HDF5	(	CudaDnn< T >	cuda,
		Log	log,
		string	strFile
	)

The constructor.

Parameters

cuda	Specifies the CudaDnn connection to Cuda.
log	Specifies the Log for output.
strFile	Specifies the HDF5 file to load.

Definition at line 45 of file HDF5.cs.

HDF5() [2/2]

MyCaffe.layers.hdf5.HDF5< T >.HDF5

(

)

The constructor used when loading weights into a Net.

Definition at line 59 of file HDF5.cs.

Member Function Documentation

CopyTrainedLayersFromHDF5()

void MyCaffe.layers.hdf5.HDF5< T >.CopyTrainedLayersFromHDF5	(	Net< T >	net,
		Log	log,
		string	strFile
	)

Copy the weights from an HDF5 file into a Net.

Parameters

net	Specifies the destination Net.
log	Specifies the output log.
strFile	Specifies the HDF5 file containing the source weights.

Implements MyCaffe.layers.hdf5.IHDF5Load< T >.

Definition at line 238 of file HDF5.cs.

Dispose()

void MyCaffe.layers.hdf5.HDF5< T >.Dispose

(

)

Release all resources uses.

Definition at line 325 of file HDF5.cs.

load_nd_dataset()

void MyCaffe.layers.hdf5.HDF5< T >.load_nd_dataset	(	Blob< T >	blob,
		string	strDatasetName,
		bool	bReshape = false,
		int	nMinDim = 1,
		int	nMaxDim = int.MaxValue,
		H5GroupId	id = null,
		bool	bAllowSingleItems = false
	)

Creates a new dataset from an HDF5 data file.

Parameters

blob	The input blob is reshaped to the dataset item shape.
strDatasetName	Specifies the new dataset name.
bReshape	Specifies whether to reshape the 'blob' parameter.
nMinDim	Specifies the minimum dimension.
nMaxDim	Specifies the maximum dimension.
id	Optional, group ID to use instead of internal file (default = null).
bAllowSingleItems	When true single item values are allowed and used to copy across entire blob.

Definition at line 170 of file HDF5.cs.

---

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

• C:/Data/Data/SS_Projects/Intelligence/GitHub/MyCaffe/MyCaffe.layers.hdf5/layers.hdf5/HDF5.cs