Support for AI Accelerator Cards

[PeterBurner]

📖 Description

Graphics cards are expensive and consume a lot of power. AI accelerator cards (M.2 or PCIe) on the other hand are often cheaper and much more power efficient. There is even an official Raspberry PI AI HAT now.

It would be nice if A-Eye could support at least some of the accelerator cards on the market.
From my research the Raspberry PI AI HAT+ and the Radxa AICore AX-M1 seem to be the most affordable.

🔎 Popular Accelerator Chips

• Hailo-8/Hailo-10 --> Raspberry PI AI HAT+, ASUS UGen300, Waveshare Hailo-8 M.2 AI Accelerator Module, Firefly Hailo-8 M.2 AI Acceleration Module
• Axera AX8850 --> Radxa AICore AX-M1, M5Stack LLM-8850, M5Stack AI Pyramid, sipeed Maix4-HAT
• MemryX MX3 --> MX3 M.2 AI Accelerator
• DeepX DX-M1 --> DeepX DX-M1 M.2 Module, Radxa AICore DX-M1
• Metis AIPU --> Metis m.2 AI Acceleration Card

🔗 Accelerator Integration in Other Projects

• Immich ML with Hailo-8
• Immich ML with Axera AX8850
• Frigate Detection with Hailo-8
• Frigate Detection with Axera AX8850
• Frigate Detection with MemryX MX3

⚙️ Unraid Drivers

• HailoRT Driver
• AXCL Host Driver
• MemryX Driver