Home

Nerlnet is a distributed machine learning platform developed for both research and deployment on real edge hardware devices.​
Nerlnet is highly configurable and very modular. It supports different kinds of communication protocols and has an easy python API for experiment management.​
The Erlang is the core of the distributed system of Nerlnet. The network of Nerlnet is based on Cowboy and it is based on common network modules: Servers, Routers and Clients. Clients can host sources that generate data in the network, and workers that execute neural networks implemented in C++, such as cppSANN, under IoT hardware constraints.