dcnnsol::feature_data_policy< CSP, PP > Struct Template Reference

Final layer activation data policy. More...

#include <dcnnelements.hpp>

Inheritance diagram for dcnnsol::feature_data_policy< CSP, PP >:

Public Types
using	values_t = tagged::permute_tensor_class< typename PP::feature_data, float, channel_tag, batch_tag>
	2-dimensional tensor type carrying activations
using	co_values_t = tagged::permute_tensor_class< tagged::co_list<typename PP::feature_data>, float, tagged::co<channel_tag>, tagged::co<batch_tag>>
	2-dimensional tensor type carrying loss derivatives wrt. activations

Static Public Attributes
static constexpr auto	cr = CSP::cr
	Channel index range, statically sized.

Detailed Description

template<typename CSP, is_policy PP>
struct dcnnsol::feature_data_policy< CSP, PP >

Final layer activation data policy.

Template Parameters

CSP	Channel sizing policy
PP	Permutation policy

Member Typedef Documentation

co_values_t

template<typename CSP, is_policy PP>

using dcnnasgn::feature_data_policy< CSP, PP >::co_values_t = tagged::permute_tensor_class< tagged::co_list<typename PP::feature_data>, float, tagged::co<channel_tag>, tagged::co<batch_tag>>

2-dimensional tensor type carrying loss derivatives wrt. activations

Same as values_t but indexes co-tagged according to tensor theory.

Layout: Indexes nested according to permutation policy (dcnnsol::permutation_policy::feature_data)

values_t

template<typename CSP, is_policy PP>

using dcnnasgn::feature_data_policy< CSP, PP >::values_t = tagged::permute_tensor_class< typename PP::feature_data, float, channel_tag, batch_tag>

2-dimensional tensor type carrying activations

Layout: Indexes nested according to permutation policy (dcnnsol::permutation_policy::feature_data)

---

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

• asgn/dcnnelements.hpp