Does DeepStream for Tesla support TensorRT Inference Server (TRTIS) to run multi models of video analytics

Accelerated Computing Intelligent Video Analytics DeepStream SDK
1

Hi all, does TensorRT Inference Server (TRTIS) support running video analytics inference with DeepStream on Tesla?. I have read that DeepStream works with TensorRT optimized inference engines as input to run the inference with DeepStream, and I need to know if DeepStream works with TRTIS also; if so, where can I find more information about it?

2

Deepstream has its own tensorRT gstreamer plugin which can set gpu_id, run multi models. You can create your pipeline including camera, decoding, inference, tracker, osd, display, … easily.

https://devblogs.nvidia.com/nvidia-serves-deep-learning-inference/ TRTIS is anohter package of tensorRT, it can support python, Kubernetes for loading balance, client/server…

They are independent from each other.

3

Hi ChrisDing, thanks for your reply.

Could you please recommend script samples that show how to work with Deepstream and tensorRT gstreamer plugin to run multi models?

4

You can download the deepstream 3.0 package. Refer to sources/apps/sample_apps/deepstream-test2