Case Study - Optimizing Linear Layer

Overview

As Uchen.ml is heading towards the public announcement and first demos, some low-hanging fruit needs to be picked in terms of optimizations. The most often used piece of any ML library is the linear layer as it is the most basic building block for any neural net. This post details the process of optimizing the code.

Requirements

Uchen is designed for implementing ML solutions that can be easily

integrated into existing systems, with specific goals on Web Assembly, embedded …

announcement basic block building case case study cpp layer library linear low machinelearning neural net overview performance process public study terms