Skip to content

Pusher

Introduction

The Pusher component pushes the model to the deployment target. The component is used in creating Training pipelines. The Robotika Pusher component is the inheritor of the Pusher TFx [1] component with an extended UX.

Inputs
  1. “Max concurrent predictions” is the upper limit of the parallel predictions. The parameter should be chosen according to the current load of your application.
  2. “Memory size” is the memory limit used for a single prediction node.
  3. “AWS account” (optional parameter), if indicated, the access would be granted to the provided AWS account by Robotika. If the parameter is not indicated, the endpoint created in Robotika will be public, so anyone with knowledge of the endpoint address will have access to it via Rest API.
References

[1] TFX pusher component by Google, https://www.tensorflow.org/tfx/guide/pusher