Common

[leondavi]

Common Questions and Answers regarding Nerlnet usage.

Q: Do I need to know Erlang to use Nerlnet?
A: Absolutely no. Nerlnet is managed through Jupyter Notebook and Json configuration files.

Q: Is it possible to run Nerlnet cluster on a single machine?
A: Yes. In DC file configure only a single device that hosts all entities.

Q: Nerlnet does not recognize my LAN IP, what should I do?
A: Under config directory go to subnets.nerlconfig and add your LAN subnet.

Q: Which OS supported?
A: Any Linux machine with gcc/g++ version 10.0 that can run Erlang-OTP version 25+

Q: Minimal machine requirements?
A: 1GB RAM and enough disk storage to run.
Use ./NerlnetBuild.sh with flag -j 1 for machines with small capacity of RAM.

Q: What infrastructure for ML on edge are supported?
A: OpenNN. There is work to add libtorch support.

Q: Is Federated ML supported?
A: Federated Average and Weighted Average are supported.