Custom Filter Class. More...
#include <tensor_filter_custom.h>
Custom Filter Class.
Note that every function pointer is MANDATORY!
Definition at line 125 of file tensor_filter_custom.h.
NNS_custom_allocate_invoke _NNStreamer_custom_class::allocate_invoke |
the main function, "allocate & invoke", that transforms input to output. allocate_invoke is supposed to allocate output buffer by itself. (invoke) XOR (allocate_invoke) MUST hold.
Definition at line 133 of file tensor_filter_custom.h.
NNS_custom_destroy_notify _NNStreamer_custom_class::destroy_notify |
it handles the data pointer allocated in the custom framework. when the data pointer has been destroyed at the pipeline, this method will be called. the data pointer or an object including data pointer could be deleted safely with this function. this method is only used when allocate_invoke is TRUE
Definition at line 134 of file tensor_filter_custom.h.
NNS_custom_exit_func _NNStreamer_custom_class::exitfunc |
will not call other callbacks after this call
Definition at line 128 of file tensor_filter_custom.h.
NNS_custom_get_input_dimension _NNStreamer_custom_class::getInputDim |
a custom filter is required to provide input tensor dimension unless setInputdim is defined.
Definition at line 129 of file tensor_filter_custom.h.
NNS_custom_get_output_dimension _NNStreamer_custom_class::getOutputDim |
a custom filter is required to provide output tensor dimension unless setInputDim is defined.
Definition at line 130 of file tensor_filter_custom.h.
NNS_custom_init_func _NNStreamer_custom_class::initfunc |
called before any other callbacks from tensor_filter_custom.c
Definition at line 127 of file tensor_filter_custom.h.
NNS_custom_invoke _NNStreamer_custom_class::invoke |
the main function, "invoke", that transforms input to output. invoke is supposed to fill in the given output buffer. (invoke) XOR (allocate_invoke) MUST hold.
Definition at line 132 of file tensor_filter_custom.h.
NNS_custom_set_input_dimension _NNStreamer_custom_class::setInputDim |
without getI/O-Dim, this allows framework to set input dimension and get output dimension from the custom filter according to the input dimension
Definition at line 131 of file tensor_filter_custom.h.