Developer Guide for Intel® Data Analytics Acceleration Library 2019 Update 4

Fully-connected Forward Layer

The forward fully-connected layer computes values

for n input arguments x k , weights w ki , weights mask s ki , and biases b i , where k ∈ {1, ..., n}, i ∈ {1, ..., m}, and m is the number of layer outputs.

Problem Statement

Given:

The problem is to compute the 2-dimensional tensor Y of size n k x m such that: