Advanced configuration

The NX AI Manager Plugin allows further integration for advanced users to add custom behavior or processing. This section covers topics such as external preprocessing and postprocessing pipelines, manual installation, plugin removal, and low-level .ini settings. These options are intended for users who need capabilities beyond the standard plugin configuration.

The diagram below shows where external preprocessors and postprocessors hook into the inference pipeline, before and after the model respectively:

Inference pipeline diagram showing four stages: Acquire Tensor, External Preprocessor (hook point), Model Inference, External Postprocessor (hook point), Output
Inference pipeline: external preprocessors run before the model, postprocessors run after

Advanced configuration topics

Last updated