Product description: reComputer J4012 - Nvidia Jetson Orin NX 16GB RAM + 128GB NVMe - Seeedstudio 110110145
The reComputer J4012 minicomputer is based on the Nvidia Jetson Orin NX 16GB, a powerful and compact intelligent edge device. Delivers up to 100 TOPS edge AI performance, offering more than 5x the performance of Jetson Xavier NX and 3x the performance of Jetson AGX Xavier. Combining the Nvidia Ampere architecture with 64-bit support, Jetson Orin NX integrates advanced multi-functional video and image processing and Nvidia Deep Learning accelerators. The device is equipped with 16 GB of RAM and 128 GB of external NVMe SSD memory.
The complete kit includes the Nvidia Jetson Orin NX 16GB module, heatsink and power supply. Nvidia JetPack 5.1 is installed on the reComputer J4011, which simplifies development and allows the implementation of Edge AI solutions working in video analytics, object detection, natural language processing, medical imaging and robotics in various industries of smart cities, security, industrial automation, smart factories .
The whole is enclosed in an aluminum housing.
reComputer J4012 is equipped with rich interfaces.
Main features of the reComputer J4012 module
- Outstanding AI performance : On-device AI performance up to 70 TOPS with low power consumption and low latency.
- Palm-sized edge AI device : compact in size 130 x 120 x 58.5mm, includes Nvidia Jetson Orin NX 8GB module, heatsink, chassis and power supply.
- Expandable through rich I/O : 4x USB 3.2, HDMI 2.1, 2x CSI, 1x RJ45 Gigabit Ethernet, M.2 Key E, M.2 Key M, CAN and GPIO.
- Fast-to-market : pre-installed Nvidia JetPack on included 128GB NVMe SSD, Linux BSP operating system, WiFi BT combo module, 2x antenna, support for Jetson software and leading AI software frameworks and platforms.
Computer Vision, Machine Vision and Speach AI
- The Nvidia DeepStream SDK provides a complete streaming analytics toolkit for AI-based multi-sensor processing and video and image understanding on Jetson.
- The Nvidia TAO Toolkit , based on TensorFlow and PyTorch, is a low-code version of the Nvidia TAO framework that speeds up model training.
- alwaysAI allows you to build, train and deploy computer vision applications directly on reComputer edge devices. It gives you free access to over 100 pre-trained computer vision models and allows you to train custom AI models in the cloud with just a few clicks with an enterprise subscription. Check out the guide .
- Edge Impulse is the easiest built-in machine learning pipeline to deploy audio, object detection classification applications
- Roboflow provides tools to convert raw images into a specially trained computer model for object classification detection and deploy the model for use in applications.
- Ultralytics' YOLOv5 : Leverage transfer learning to realize object detection with YOLOv5, which requires only a few training samples. Check out the guide .
- Deci : Optimize your models on Nvidia Jetson Nano. Check out this webinar at Deci on automated benchmarking and runtime performance optimization on Nvidia Jetson Nano and Xavier NX devices.
- Nvidia Riva is a GPU-accelerated SDK for building Speach AI apps that are tailored to your use case and deliver real-time performance.
The reComputer J4012 comes with a power adapter ( power cord must be purchased separately).
Jetson precompile system.
Pre-installed Nvidia JetPack
The Nvidia JetPack SDK provides a complete development environment and includes CUDA-X-accelerated libraries and other Nvidia technologies to accelerate development. JetPack includes the Jetson Linux Driver Pack which provides the Linux kernel, boot loader, Nvidia drivers, flashing tools, sample file system and toolchains for the Jetson platform. It also includes security features, over-the-air upgradeability, and much more.
Kit contents
- Nvidia Jetson Orin NX 16GB module
- Seeed carrier plate
- 128GB NVMe SSD
- Aluminum radiator with fan
- Aluminum housing (black)
- Power adapter 12 V / 5 A (DC 5.5 / 2.5 mm, power cord not included)
Interfaces in reComputer Jetson
The J4012 is a small board with a rich interface and high performance that allows you to bring new possibilities to all embedded AI and edge systems.
- Control and UART connector
- RTC 2-pin connector
- 2x MIPI-CSI camera connector
- 4x USB 3.0 ports
- HDMI and DisplayPort ports
- DC jack power connector
- CAN port
- Fan connector
- 40-pin GPIO connector
- 260-pin SO-DIMM connector
- PoE connector
- Gigabit Ethernet port
- Led
- USB Type-C
- M.2 Key E connector
- RTC coin cell battery basket
- M.2 Key Connector M
Pins of the reComputer J401x board.
| SPECIFICATION AND COMPARISON | ||
|---|---|---|
| Product | reComputer J4011 | reComputer J4012 |
| Module | Jetson Orin Nano 8GB | Jetson Orin Nano 16GB |
| AI performance | 700 TOPS | 700 TOPS |
| GPU | ||
| GPU clock | 765MHz | 918MHz |
| CPU | 6 Core Nvidia Arm Cortex A78AE v8.2 1.5MB L2 + 4MB L3 | 8 Cores Nvidia Arm Cortex A78AE v8.2 2MB L2 + 4MB L3 |
| CPU clock speed | ||
| RAM | 8GB 128-bit LPDDR5 102.4GB/s | 16GB 128-bit LPDDR5 102.4GB/s |
| DL accelerator | 1x NVDLA v2 | 2x NVDLA v2 |
| Maximum DLA frequency | ||
| Vision Accelerator | ||
| Memory | ||
| video encoder | 1x 4K60 (H.265) | 3x 4K30 (H.265) 6x 1080p60 (H.265) | 12x 1080p30 (H.265) |
|
| Video decoder | 1x 8K30 (H.265) | 2x 4K60 (H.265) | 4x 4K30 (H.265) 9x 1080p60 (H.265) | 18x 1080p30 (H.265) |
|
| display | ||
| CSI camera | ||
| Network | ||
| USB | 1x USB 2.0 Type-C (Device Mode) |
|
| M.2 Key M | ||
| M.2 Key E | ||
| Fan connector | ||
| CAN | ||
| Multi-functional port | ||
| RTC | ||
| Supply voltage | ||
| Power | ||
| Working temperature | ||
| Dimensions | ||
