how to use created .engine file for inference


I created mnist.engine file using this sample to have the understanding of how tensorrt works using this link

I am not able to understand hoe to use this mnist.engine file for inference.

My ideal point of learning to use tensorrt is I want to create inference engine for object detection using Faster RCNN architecture and Caffe framework with Titan X can I use Tensorrt for this?

I am not getting intuitive understanding.

is dataset needed for inference?
what is binaryproto file what’s the use of it?

i am completely into python so any explanation in python perspective. will be more helpful for me.

can anyone help me please.
Thank you in advance .

Here are a couple of resources you might find helpful
(in particular the section labeled “TensorRT run time inference”)

Hope this helps!

We also have a deep learning institute with online and face to face courseware to help you get an intuitive understanding of this.

For example the “Fundamentals of Deep Learning for Computer Vision” course includes object detection and deployment with TensorRT.

This Fundamentals course will give the developer a better understanding of how to implement and deploy deep learning capabilities in their applications:

Additional self-paced training available at

Also, it is kind of Jetson specific but might still be interesting: