Build a Triton Inference Server with Preprocess and Postprocess: Part (2/4)

Author(s): Hao WEI

Originally published on Towards AI.

A step-by-step coding practice

a step-by-step coding practice

pub.towardsai.net

In my last blog, we built a bare minimum triton server using an MNIST example. In the example, we did preprocessing and postprocessing in the client side. In the real world, preprocessing and postprocessing may be complicated or computationally heavy.

preprocess : In the CV domain, preprocessing includes normalization, resizing, orienting, color correction, etc. In the NLP domain, it includes tokenization, stemming, stop-word removal, etc.postprocess: In the CV domain, Non-Maximum Suppression (NMS), for example. In NLP, you need to decode the softmax output to text class names, for example.

The two modules may need some specific libraries to run… Read the full blog for free on Medium.

Join thousands of data leaders on the AI newsletter. Join over 80,000 subscribers and keep up to date with the latest developments in AI. From research to projects and ideas. If you are building an AI startup, an AI-related product, or a service, we invite you to consider becoming a sponsor.

Published via Towards AI