DeepStream SDK 6.2 Release Notes

Discover the latest advancements in NVIDIA's DeepStream SDK with version 6.2, designed for NVIDIA dGPU/X86 and Jetson platforms.

About This Release

This release introduces significant updates and new features for the DeepStream SDK, enhancing capabilities for AI and intelligent video analytics applications. Key highlights include support for Triton 22.09 and 23.01, improved DeepSORT tracker, REST API integration, and LIDAR support.

What's New

  • Enhanced support for Triton inference server versions.
  • Introduction of DeepSORT tracker and REST API for pipeline control.
  • Added LIDAR support and enhancements for dewarper functionality.
  • New nvdsxfer plugin for multi-GPU communication via NVLink.
  • Performance optimizations and improved trackers like NVDCF.
  • GPU-based drawing capabilities for visual elements.
  • Integration with NVIDIA TAO toolkit models.
  • New sample applications demonstrating various features like Lidar inference and Can orientation detection.

Key Features and Updates

  • DS 6.2: Features support for the latest Triton versions and Jetson package updates.
  • Graph Composer 2.5: Offers an improved UI, graph runtime, and tools for building and managing AI pipelines.
  • Documentation: Comprehensive guides are available, including the Developer Guide, API Reference, and Python API Reference.

Important Notes and Limitations

  • Details on limitations such as V4L2 codec instance limits and specific configuration requirements are provided.
  • Information on deploying applications in Docker containers and integrating with Azure IoT Edge is included.
  • Guidance on installing DeepStream on Jetson and using Triton Inference Server with various frameworks like TensorRT, TensorFlow, ONNX, and PyTorch is available.

PDF preview unavailable. Download the PDF instead.

DeepStream 6.2 Release Notes Adobe PDF Library 22.3.90

Related Documents

Preview NVIDIA DeepStream SDK 6.4 Release Notes for DGPU/X86 and Jetson
This document provides release notes for NVIDIA DeepStream SDK 6.4, detailing new features, changes, limitations, and notes for DGPU/X86 and Jetson platforms. It covers updates to DS 6.4, DS 6.3, Graph Composer 3.1.0, and breaking changes from previous versions.
Preview NVIDIA DeepStream SDK 6.0 Release Notes for DGPU and Jetson
This document provides release notes for NVIDIA DeepStream SDK version 6.0, detailing new features, enhancements, limitations, and important notes for developers working with NVIDIA DGPU and Jetson platforms. It covers updates to DS 6.0, Graph Composer 1.0.0, and integration with Triton Inference Server.
Preview NVIDIA DeepStream SDK 7.1 Release Notes
This document provides release notes for NVIDIA DeepStream SDK 7.1, detailing new features, improvements, limitations, and deprecations for NVIDIA DGPU/X86 and Jetson platforms.
Preview NVIDIA DeepStream 4.0 Plugin Manual
A comprehensive guide to NVIDIA DeepStream SDK's GStreamer plugins, detailing their functionality, inputs, outputs, features, and configuration for AI-powered video analytics and computer vision applications.
Preview NVIDIA Jetson Nano: Hello AI World - Your Guide to AI at the Edge
Explore the NVIDIA Jetson Nano, a powerful and accessible platform for AI and robotics development. This guide covers its specifications, software stack (JetPack, TensorRT), application SDKs (DeepStream, Isaac), getting started tutorials, and use cases for autonomous machines.
Preview NVIDIA Jetson Linux 34.1.1 Developer Preview Release Notes
This document provides release notes for NVIDIA Jetson Linux 34.1.1 Developer Preview, detailing new features, known issues, fixed issues, and implementation details for NVIDIA Jetson platforms including AGX Orin and Xavier NX.
Preview NVIDIA Triton Management Service Deployment Guide
This guide provides instructions for deploying and managing NVIDIA Triton Inference Servers using the Triton Management Service (TMS) within a Kubernetes cluster. It covers basic operations, lease management, and inference execution.
Preview NVIDIA AI Enterprise User Guide: Installation, Configuration, and Management
Comprehensive user guide for NVIDIA AI Enterprise, detailing installation, configuration, and management of NVIDIA vGPU, AI frameworks, and software components across various hypervisors and operating systems.