Deep Neural Network Library (DNNL)  1.1.3
Performance library for Deep Learning
List of all members
dnnl::concat Struct Reference

Implements primitive descriptor and primitive for concat. More...

#include <dnnl.hpp>

Inheritance diagram for dnnl::concat:
Inheritance graph
[legend]
Collaboration diagram for dnnl::concat:
Collaboration graph
[legend]

Additional Inherited Members

- Public Types inherited from dnnl::primitive
enum  kind {
  kind::undef = dnnl_undefined_primitive, kind::reorder = dnnl_reorder, kind::shuffle = dnnl_shuffle, kind::concat = dnnl_concat,
  kind::sum = dnnl_sum, kind::convolution = dnnl_convolution, kind::deconvolution = dnnl_deconvolution, kind::eltwise = dnnl_eltwise,
  kind::softmax = dnnl_softmax, kind::pooling = dnnl_pooling, kind::lrn = dnnl_lrn, kind::batch_normalization = dnnl_batch_normalization,
  kind::layer_normalization = dnnl_layer_normalization, kind::inner_product = dnnl_inner_product, kind::rnn = dnnl_rnn, kind::binary = dnnl_binary
}
 Kinds of primitives. More...
 
- Public Member Functions inherited from dnnl::primitive
const_dnnl_primitive_desc_t get_primitive_desc () const
 Returns the descriptor of the underlying C API primitive.
 
- Public Member Functions inherited from dnnl::handle< dnnl_primitive_t >
 handle ()=default
 Empty constructor. More...
 
 handle (dnnl_primitive_t t, bool weak=false)
 Constructs a C handle wrapper from a C handle. More...
 
void reset (dnnl_primitive_t t, bool weak=false)
 Resets the value of a C handle. More...
 
dnnl_primitive_t get (bool allow_emtpy=false) const
 Returns the value of the underlying C handle.
 

Detailed Description

Implements primitive descriptor and primitive for concat.

Creates an out-of-place primitive descriptor for concatenation of n inputs by concat_dimension with resulting output_desc memory descriptor. output_desc can be NULL or specified with the dnnl::memory::format_tag::any format kind–in this case, the appropriate memory format would be chosen automatically.


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