reComputer J1010 - Nvidia Jetson Nano 4GB RAM + 16GB eMMC - Seeedstudio 110061362

Product description: reComputer J1010 - Nvidia Jetson Nano 4 GB RAM + 16 GB eMMC - Seeedstudio 110061361

The minicomputer reComputer from the Jetson series is based on the Nvidia Jetson Nano module with advanced artificial intelligence (AI) systems with 4 GB of RAM and 16 GB of eMMC memory . With a range of expansion modules and industrial peripherals built into the board, as well as thermal management, Jetson reComputer is ready to help accelerate and scale up next-generation AI products by deploying popular DNN models and ML frameworks on the edge and high-performance inference for tasks such as real-time object classification and detection, position estimation, semantic segmentation and natural language processing (NLP). Developers can use pre-trained models from Tao Toolkit and deploy them via DeepStream. Jetson Nano can achieve 11 FPS for PoepleNet - ResNet33 for human recognition, 19 FPS for DashCamNet-ResNet18 for vehicle recognition, and 101 FPS for FaceDetect-IR-ResNet18 for facial recognition. Test results can be found on NVIDIA®'s DeepStream SDK website .

Jetson Nano allows you to run a wide variety of advanced networks, including full native versions of popular machine learning (ML) frameworks such as TensorFlow, PyTorch, Caffe / Caffe2, Keras, MXNet and others.

reComputer Jetson is compatible with the entire Nvidia Jetson software stack, industry-leading AI platforms, development platforms like Edge Impulse , Always AI , and development tools for Nimbus cloud-based robots.

reComputer J1010.

Main features of reComputer J1010

• Compact AI edge device based on Nvidia Jetson Nano module, offering 128 Nvidia Cuda cores providing 0.5 TFLOPs (FP16).
• Twin with Nvidia Jetson Nano Debeloper Kit, many I / O pins including Gigabit Ethernet port, USB 3.0 port and USB 2.0 type A and HDMI port.
• Official Nvidia JetPack software pre-installed, ready for cloud-native applications.
• It supports a wide range of fast-growing AI applications.
• Supports real-time robot development, analysis and remote fleet management in Cogniteam Nimbus .
• It supports Allxon to enable secure OTA and remote device management. 90 day trial period with the code: H4U-NMW-CPK.
• Can be mounted on the wall thanks to the mounting holes.

Application of reComputer Jetson

The reComputer Jetson series is ideally suited to the construction of autonomous machines and complex AI system tasks in the field of image recognition, object detection and location, movement recognition, semantic segmentation, video processing and intelligent analysis.

Edge AI into the Wild : UAV, Drone, Wildlife Conservation, Agriculture

Smart cities : traffic, retail, healthcare

Industry : production, logistics, delivery, service

Jetson precompilation system.

Pre-compilation system for Edge AI integration

Nvidia Jetson series modules are supported by the same software stack, so once written application can be run on other Jetson boards. Jetson software is designed to provide end-to-end acceleration for AI applications and reduce time-to-market. It includes the same advanced Nvidia technologies found in data centers and cloud-based edge deployments. Nvidia JetPack includes Nvidia Container Runtime with Docker integration to enable GPU acceleration of containerized applications on the Jetson platform. Jetpack also supports Nvidia Triton Inference Server to simplify the large-scale deployment of AI models.

Kit contents

• Acrylic cover
• Aluminum frame
• Jetson Nano module
• Radiator
• Carrier plate

Interfaces in reComputer Jetson

On the board compatible with Nvidia Jetson there are ports such as HDMI 2.0, Gigabit Ethernet, USB 3.0 and USB 2.0 port, CSI, GPIO, I2C, I2S camera connectors, fan connector and other rich peripheral interfaces. It has almost the same functional design and size as the Nvidia Jetson Nano board. Small size, multitude of interfaces and high performance allow to introduce new possibilities to all embedded artificial intelligence and edge systems.

1. MIPI-CSI camera connectors
2. USB Type-C power connectors
3. Type A HDMI port
4. USB 3.0 Type A
5. 2x USB 2.0 type A connectors
6. Gigabit Ethernet
7. USB Type-C (2.0) connector
8. Button connector
9. MicroSD card slot (reserved)
10. M.2 Key E connector
11. 2-pin RTC connector
12. 40-pin connector
13. 260-pin (SO-DIMM) connector
14. Fan connector

Pins of the Jetson reComputer board.

Specification and comparison Product reComputer JJ1010 reComputer J1020 Nvidia Jetson Nano-B01 Module Nano (development version) Nano (non-developer version) AI performance 472 GFLOPS GPU 128-core Nvidia Maxwell chip Processor Quad-Core ARM A57 1.43 GHz RAM 4 GB 64-bit LPDDR4 25.6 GB / s Internal storage 16 GB eMMC MicroSD Video codec 4K30 | 2x1080p60 | 4x1080p30 | 4x720p60 | 9x720p30 (H.265 and H.264) 4Kp30 | 4x 1080p30 | 9x 720p30 (H.264 / H.265) Video decoder 4K60 | 2x 4K30 | 4x 1080p60 | 8x 1080p30 | 9x 720p60 (H.265 and H.264) 4Kp60 | 2x 4Kp30 | 8x 1080p30 | 18x 720p30 (H.264 / H.265) Ethernet connection 1x RJ45 Gigabit Ethernet (10/100/1000) USB

1x USB 3.0 type A connector 2x USB 2.0 Type A 1x USB Type-C host connector 1x USB Type-C connector for power supply

4x USB 3.0 type A connector 1x microUSB port host 4x USB 3.0 type A connector 1x microUSB port for power and host Camera connector 2x CSI (15-pin, 1 mm pitch, MIPI-CSI-2) Display connector 1x HDMI type A 1x HDMI type A 1x DP

1x HDMI type A 1x DP

Cooling 1x fan connector M.2 key E 1x M.2 key E 1x M.2 E key (disabled) 1x M.2 key E M.2 key M - 1x M.2 M key - RTC 1x RTC socket - GPIO 1x 40-pin Power USB type C 5V / 3A 12 V / 2 A jack connector DC 5V / 3A jack microUSB 5 V / 2 A Dimensions 130 x 120 x 50 mm (with housing) 130 x 120 x 50 mm (with housing) 100 x 80 x 29 mm

Useful links

• The website of the manufacturer: SeeedStudio 110061362
• ReComputer Jetson User Guide
• Maskcam User Guide
• Allxon User Guide