NVIDIA Jetson Orin Nano Developer Kit sets a new standard for creating entry-level AI robots, intelligent drones, and intelligent vision systems, while simplifying the entry into the NVIDIA Jetson Orin Nano series. This developer kit’s compact design, numerous connectors, and AI performance of up to 40 TOPS make it ideal for turning your visionary concepts into reality.
This developer kit consists of a Jetson Orin Nano 8GB module and a reference carrier board that accommodates all NVIDIA Jetson Orin Nano and NVIDIA Jetson Orin NX modules, providing an ideal platform for prototyping next-generation edge AI products.
The Jetson Orin Nano 8GB module features an NVIDIA Ampere architecture GPU with 1024 CUDA cores, 32 third-generation Tensor Cores, and a 6-core Arm CPU, enabling multiple concurrent AI application pipelines and high-performance inference. The developer kit carrier board has a wide array of connectors, including two MIPI CSI connectors, supporting up to four channels of camera modules, achieving higher resolutions and frame rates than before.
The previous generation Jetson Nano developer kit made AI accessible. The new Jetson Orin Nano developer kit raises the entry-level AI development bar with an 80x increase in performance, empowering developers to run any type of modern AI model, including Transformers and advanced robotic models. Compared to the previous generation Jetson Nano, the Jetson Orin Nano not only significantly improves AI performance but also offers 5.4x CUDA compute capability, 6.6x CPU performance, and 50x performance per watt.
Figure 1 Performance and efficiency comparison of NVIDIA Jetson Orin Nano and NVIDIA Jetson Nano
NVIDIA Jetson Orin Nano Developer Kit Features
Figure 2 The Jetson Orin Nano Developer Kit includes a Jetson Orin Nano 8GB compute module, reference carrier board, heatsink/fan, DC power supply, and 802.11ac / abgn WLAN + BT module
Tables 1 and 2 show the main features and interfaces of the kit carrier board.
Table 1 NVIDIA Jetson Orin Nano Developer Kit Module Description
Table 2 NVIDIA Jetson Orin Nano Developer Kit Carrier Board Specifications
The Jetson Orin Nano Developer Kit is now available for pre-order and will begin shipping in April. For Jetson Orin Nano technical documentation, reference design files, and software, please visit Jetson Download Center (https://developer.nvidia.com/embedded/downloads) .
Jetson Orin Nano Runs All Modern AI Models
The Jetson Orin Nano Developer Kit offers up to 40 TOPS of AI performance, capable of running all modern AI models. This significant leap in computation enables the most demanding AI applications, including running Transformer models at the edge, which was previously impossible.
Transformer models are the foundation of recent generative AI applications, such as ChatGPT and DALL-E, which are sweeping the world. Transformer models learn context and meaning by tracking relationships between elements in sequential data without the need for large labeled datasets.
Start today:
-
City Segmentation (https://catalog.ngc.nvidia.com/orgs/nvidia/teams/tao/models/citysemsegformer) for segmenting urban landscapes into different categories
-
PeopleNet Transformer (https://catalog.ngc.nvidia.com/orgs/nvidia/teams/tao/models/peoplenet_transformer) based on Deformable Detection Transformer
-
BERT (https://docs.nvidia.com/deeplearning/riva/user-guide/docs/reference/models/nlp.html#bert) for natural language processing
Jetson Software Accelerates AI and TTM
The Jetson Orin Nano Developer Kit features use-case-specific available application frameworks that run the NVIDIA AI software stack. This includes NVIDIA Isaac for robotics, NVIDIA DeepStream for visual AI, and NVIDIA Riva for conversational AI. Using NVIDIA Omniverse Replicator for synthetic data generation (SDG) can save significant time, and using NVIDIA TAO Toolkit also allows fine-tuning pre-trained AI models from the NGC catalog.
Figure 3 shows the results of running some computer vision benchmarks based on Jetson Orin Nano using the upcoming NVIDIA JetPack 5.1.1. These results indicate that the developer kit raises the standard for entry-level computer vision. The tests include some intensive INT8 and FP16 pre-trained models from NGC, as well as the Industry Resnet-50 benchmark. The benchmarks include the following:
-
NVIDIA PeopleNet v2.5 for maximum accuracy in person detection
-
NVIDIA ActionRecognitionNet for 2D and 3D models
-
NVIDIA LPRNet for license plate recognition
-
for multi-person pose estimation NVIDIA DashCamNet and BodyPoseNet
-
ResNet-50 (224 × 224) object detection model
The Jetson Orin Nano Developer Kit is a versatile platform that supports models trained with NVIDIA TAO Toolkit 4.0 and will soon support newly released models in TAO Toolkit 5.0. With TAO Toolkit 5.0, developers can leverage several state-of-the-art visual Transformer models for image classification, object detection, and segmentation use cases. For more information, see:
https://developer.nvidia.com/blog/access-the-latest-in-vision-ai-model-development-workflows-with-nvidia-tao-toolkit-5-0/
NVIDIA Jetson Orin Nano and NVIDIA DeepStream are the ideal combination for edge applications such as smart retail, smart city intersections, and industrial automation. With the upcoming version of DeepStream and the introduction of GXF runtime, Jetson Orin Nano will be the ideal platform for running AI graphs that need to tightly integrate with deterministic systems, which are common in factory automation use cases.
Additionally, for those looking to quickly familiarize themselves with DeepStream, you can use the latest version of Jetson Graph Composer to build applications and deploy them to Jetson Orin Nano with a click of a button.
Figure 4 Build applications in DeepStream Graph Composer and deploy to Jetson Orin Nano
Using NVIDIA Isaac on NVIDIA Jetson Orin to Accelerate Robotics Applications
NVIDIA Isaac robot platform is a powerful end-to-end platform for developing, simulating, and deploying AI robotics. The hardware acceleration package NVIDIA Isaac ROS makes it easier for ROS 2 developers to build high-performance solutions on the Jetson Orin Nano Developer Kit. The new NVIDIA Isaac ROS DP optimizes the ROS 2 node processing pipeline to execute on the Jetson Orin platform. It also provides new DNN-based GEMS designed to improve throughput.
Figure 4 shows the results of running these robotics packages on Jetson Orin Nano using the upcoming NVIDIA Isaac ROS DP3. The performance is measured under load, including message transmission costs in RCL, indicating real-world performance benchmarks. The tests include algorithms shown in the figure (Figure 4), including:
-
Visual SLAM enables robots to compute their position and movement based on images by tracking visual features around the environment
-
April Tags for AprilTag detection and pose estimation
-
Image Detection
-
Image Segmentation
-
Proximity Segmentation to determine if obstacles are within close range and avoid collisions during navigation
-
Stereo Disparity for capturing stereo input images and generating disparity maps for robotic navigation
Learn how to get started with the NVIDIA Jetson Orin Nano Developer Kit:
Start building next-generation application prototypes, pre-order the NVIDIA Jetson Orin Nano Developer Kit (https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/jetson-orin/) and install the latest NVIDIA JetPack.
Figure 6 NVIDIA Jetson Orin Nano Developer Kit
The entire line of NVIDIA Jetson Orin series modules is now available for mass production, including the Jetson AGX Orin series, Jetson Orin NX series, and Jetson Orin Nano series. The Jetson AGX Orin Developer Kit has been upgraded to include 64GB of memory.
Additional Resources:
For more information about the NVIDIA Jetson Orin Nano Developer Kit, watch the NVIDIA GTC23 session:Connect with the Experts: A Deep-Dive Q&A on Jetson Orin and Edge AI (https://www.nvidia.com/gtc/session-catalog/?tab.catalogallsessionstab=16566177511100015Kus&search=CWES52132#/session/1670472065129001rf2z).
To learn more about this module, see Solving Entry-Level Edge AI Challenges with NVIDIA Jetson Orin Nano (https://developer.nvidia.com/blog/solving-entry-level-edge-ai-challenges-with-nvidia-jetson-orin-nano/#:~:text=Jetson%20Orin%20Nano%20series%20%20%20%20,CPU%20%20…%20%2014%20more%20rows%20) and Jetson Orin Nano Series Data Sheet (https://developer.nvidia.com/embedded/downloads#?search=Jetson%20Orin%20Nano%20Series%20Module%20Data%20Sheet).
Visit the NVIDIA Embedded Developer page (https://developer.nvidia.com/embedded-computing) and forums (https://forums.developer.nvidia.com/c/agx-autonomous-machines/jetson-embedded-systems/70) for help from community experts.
Scan the QR code on the poster, or click “Read the original text” to watch the replay of NVIDIA founder and CEO Jensen Huang’s GTC23 keynote speech!
