View HardwareDownload PDF

 

Lanner partners with DEEPX to deliver high-performance computer vision for smart city applications. Powered by the Lanner EAI-I510X, featuring the latest Intel® Core™ Ultra processors and 10GbE networking, the solution integrates the DEEPX DX-M1 NPU to enable energy-efficient, high-throughput AI inference. Deployment is further simplified with the DEEPX DXNN® Software Development Kit, providing an integrated, version-aligned toolchain for AI model compilation, optimization, simulation, and inference—accelerating time to deployment at the edge.


 

DEEPX DX-M1

DEEPX DX-M1 is installed into the host’s M.2 2280 slot, adding powerful AI inference capabilities (25 TOPS) with high power efficiency (Max 5W).

Thermal Efficiency

DEEPX chips uniquely maintain high performance within the industrial temperature range (-25°C to 85°C), ensuring consistent operation without thermal throttling in edge environments.

GPU-Level of AI Accuracy

DEEPX achieves GPU-level accuracy using optimized Int-8 quantization. Unlike competitors, it strictly meets the <1% accuracy loss standard for consistent and reliable predictions.

Unmatched AI Performance

DEEPX delivers 240% higher performance and 20x better power efficiency than GPUs. This ensures reliable AI processing in power-constrained environments without degradation.

Industry-Leading TCO

DX-M1 saves ~94% in electricity costs over five years vs GPGPUs. It remains more economical than free competitor chips by drastically reducing power and cooling requirements.

 

DXNN SDK Full Stack Architecture

DXNN® (DEEPX Neural Network) SDK streamlines AI deployment on DEEPX NPUs by integrating version-aligned tools for compilation, optimization, simulation, and inference. For efficient development, it’s offered as the DX-AS (AllSuite), a fully integrated and optimized package.


Featured Product
 

 

EAI-I500X Edge AI Computer

  • CPU: Intel® Core™ Ultra 7 165H
  • Memory: DDR5 32GB
  • Storage: SSD 500GB
  • 2x 2.5GbE RJ45, 2x 10GbE RJ45
  • SOM TDP: Up to 45W / 28W

 

DEEPX DX-M1 M.2

  • 25 TOPS (INT8 Performance)
  • PCIe Gen.3 x4
  • M.2 2280 (Key M)
  • 2W (min) ~ 5W (max)