BufferStore Class Reference

Interface for object which can store buffers. More...

Detailed Description

Interface for object which can store buffers.

#include <BufferStore.h>

Inheritance diagram for BufferStore:

Public Member Functions
	BufferStore (const OptionSet &options, NEML2Object *object)
const Storage< std::string, TensorValueBase > &	named_buffers () const
Storage< std::string, TensorValueBase > &	named_buffers ()
template<typename T , typename = typename std::enable_if_t<std::is_base_of_v<BatchTensorBase<T>, T>>>
T &	get_buffer (const std::string &name)
	}@
virtual void	send_buffers_to (const torch::TensorOptions &options)
	Send all buffers to options.
template<typename T , typename = typename std::enable_if_t<std::is_base_of_v<BatchTensorBase<T>, T>>>
const T &	declare_buffer (const std::string &name, const T &rawval)
	Declare a buffer.
template<typename T , typename = typename std::enable_if_t<std::is_base_of_v<BatchTensorBase<T>, T>>>
const T &	declare_buffer (const std::string &name, const std::string &input_option_name)
	Declare a buffer.

Constructor & Destructor Documentation

BufferStore()

BufferStore	(	const OptionSet &	options,
		NEML2Object *	object )

Member Function Documentation

declare_buffer() [1/2]

template<typename T , typename >

const T & declare_buffer ( const std::string & name, const std::string & input_option_name )

protected

Declare a buffer.

Note that all buffers are stored in the host (the object exposed to users). An object may be used multiple times in the host, and the same buffer may be declared multiple times. That is allowed, but only the first call to declare_buffer constructs the buffer value, and subsequent calls only returns a reference to the existing buffer.

Template Parameters

T	Buffer type. See Statically shaped tensors for supported types.

Parameters

name	Buffer name
input_option_name	Name of the input option that defines the value of the model buffer.

Returns

T Reference to buffer

declare_buffer() [2/2]

template<typename T , typename >

const T & declare_buffer ( const std::string & name, const T & rawval )

protected

Declare a buffer.

Note that all buffers are stored in the host (the object exposed to users). An object may be used multiple times in the host, and the same buffer may be declared multiple times. That is allowed, but only the first call to declare_buffer constructs the buffer value, and subsequent calls only returns a reference to the existing buffer.

Template Parameters

T	Buffer type. See Statically shaped tensors for supported types.

Parameters

name	Buffer name
rawval	Buffer value

Returns

Reference to buffer

get_buffer()

template<typename T , typename >

T & get_buffer

(

const std::string &

name

)

}@

Get a writable reference of a buffer

named_buffers() [1/2]

Storage< std::string, TensorValueBase > & named_buffers

(

)

named_buffers() [2/2]

const Storage< std::string, TensorValueBase > & named_buffers ( ) const

inline

Returns

the buffer storage

send_buffers_to()

void send_buffers_to ( const torch::TensorOptions & options )

protectedvirtual

Send all buffers to options.

Parameters

options

The target options