The Raspberry Pi 4 - AKD1000 Dev Kit is an all-in-one development platform combining a Raspberry Pi 4 with BrainChip's Akida AKD1000 neuromorphic processor. Enables low-power neural network inference directly on device for edge AI prototyping and evaluation.
What's in the box
- Raspberry Pi 4-based device
- AKD1000 PCIe Board (pre-assembled)
- Pre-configured 32GB SD Card (Akida demos + MetaTF)
- 12V Power Supply
- Quick-Start Guide
- Screwdriver
Specifications
- SoC: Broadcom BCM2711C0, quad-core Arm Cortex-A72 (ARMv8-A) 64-bit @ 1.5GHz
- GPU: Broadcom VideoCore VI
- RAM: 8GB LPDDR4 SDRAM
- Networking: Gigabit Ethernet; optional Wi-Fi
- Akida: AKD1000 PCIe board (pre-installed)
- Interfaces: PCIe, 2× DSI, 2× CSI, 2× HDMI
- Storage: External MicroSD (CM4 Lite); 8/16/32GB eMMC (CM4)
- Ports: Hirose U.FL antenna connector, 2× 100-pin connectors
- Dimensions: 20.06cm × 10.5cm × 3.7cm (exc. carrier board)
- OS: Linux (MetaTF pre-installed)
Designed for
- Edge AI Prototyping
- Visual Object Classification
- Anomaly Detection
- Edge Impulse Model Deployment
Quick Start
Wireless (Wi-Fi units only)
- Connect power supply
- Connect to 'akida-devkit' Wi-Fi on your phone/laptop
- Navigate to http://10.42.0.1 in a browser (popups enabled)
Wired
- Connect power supply
- Connect device to a network switch/router with an ethernet cable
- Navigate to http://<device_ip_address> in a browser (popups enabled). IP address assigned by your DHCP server.
