Frequently Asked Questions#

Who is OctaiPipe built for ?#

OctaiPipe is an end-to-end open source platform for deploying AI and ML to the edge. It has a comprehensive, flexible architecture that enables both data science experts and executives to accelerate the journey towards scalable AI at the edge.

How can I optimize cloud cost of running OctaiPipe?#

1. Select suitable size for your Notebook servers. Start with a small server, like 1 CPU and 2 GB of memory. It should be enough to orchestrate your OctaiPipe workflows. If you find that you need more to do some analysis on the server - slowly iterate sizes from there.

2. Make sure your stop your Notebook servers when you are not using them. Simply click ‘stop’ symbol against the server you want to stop and turn it on again when resuming work, otherwise underlying k8s node will be running up the tab!

3. Do as much training on the devices as you can. Using FL is cost-saving in itself because you are not running expensive cloud servers.