DFRobot DX-M1 AI Accelerator M.2 Module

25 TOPS | 4GB LPDDR5 | PCIe Gen3 x4 🚀

Please note: This is an M.2 M-Key 2280 module. For Raspberry Pi 5 use, a compatible M.2 HAT or adapter is required (not included).Key FeaturesEdge AI Power

25 TOPS (INT8) of inference performance for object detection, video analytics and visual SLAM, with no cloud needed.

Low Power

Runs within a 2W to 5W power range at 3.3V, ideal for robots and embedded systems.

Standard M.2 2280

M.2 M-Key, PCIe Gen3 x4 (compatible with x1 mode) for x86 PCs, LattePanda and Raspberry Pi 5 setups.

Developer Friendly

Supported by the DXNN SDK with PyTorch, ONNX, TensorFlow, TensorFlow Lite, Keras and XGBoost support. ⚙️

SpecificationsPerformance25 TOPS (INT8)InterfaceM.2 M-Key, PCIe Gen3 x4 (x1 compatible)Memory4GB LPDDR5, 1Tbit QSPI NAND FlashPower Consumption2W to 5WSupply Voltage3.3V ±5%Operating SystemsWindows 10/11, Ubuntu 20.04/22.04 LTSOperating Temperature-25°C to 85°C (throttling); 25°C to 65°C (non-throttling)Dimensions22 x 80 x 4.1 mm (0.87 x 3.15 x 0.16 in)Applications

Package Contents

PhotosFAQDoes it work with Raspberry Pi 5?

Yes, through a compatible M.2 HAT or adapter, which is sold separately.

Which operating systems are supported?

Windows 10/11 and Ubuntu 20.04/22.04 LTS.

Where can I find the SDK and documentation?

Message us and we will help you find the official SDK, model zoo and quick start guides from the manufacturer.