Skip to content

About

Builds, binaries, tutorials for OrangePi 4A AllWinner device

Resources

Stars

6 stars

Watchers

2 watching

Forks

Latest commit

 

History

93 Commits

Folders and files

NameName
Last commit message
Last commit date
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Builds, binaries, tutorials for OrangePi 4A AllWinner device

Hardware Acceleration & NPU Modules

  • NPU Modules for Linux 6.18.44 (NPU_modules_opi4a_6.18.44): Production activation suite for the VeriSilicon VIP9000 NPU on Allwinner T527 (Orange Pi 4A). Includes patched etnaviv.ko kernel module, out-of-tree source, patched Mesa Teflon delegate (libteflon.so), on-demand NPU controller (NPU_enable.sh), dynamic telemetry dashboard (opi-mon), and automated installer (install.sh).
  • Vendor VIPCore & VIPLite NPU Modules (NPU_modules_opi4a_6.18.44_vendor): Complete vendor NPU driver suite and dual-stack management framework for the Allwinner T527 VIP9000. Features the ported VIPCore v1.13 kernel driver (vipcore.ko), CCU power domain & clock management integration, userspace VIPLite v1.13 runtime (libVIPlite.so), native T527 NBG model runner (vpm_run), real-time dual-mode telemetry dashboard (opi-mon), and the canonical unified driver lifecycle manager (npu / npu-switch).
  • NPU Applications Suite for Linux 6.18.44 (NPU_proj_opi4a_6.18.44): Production edge AI applications for the Orange Pi 4A (Allwinner T527 VIP9000). Includes sub-millisecond thermodynamic neural surrogates (IAPWS-97 steam tables), real-time vision AI (MobileNet V1 classification & SSD MobileNet object detection), and edge face clustering & identity indexing (SCRFD + MobileFaceNet + SQLite).
  • Vendor VIPCore NPU Applications Suite (NPU_proj_opi4a_6.18.44_vendor): Production deep learning applications utilizing the vendor VIPCore driver (/dev/vipcore) and VIPLite v1.13 runtime on the Allwinner T527 VIP9000 NPU. Includes industrial two-stage layout-guided document OCR (PicoDet-Layout + PP-OCRv6 DBNet at 1600x1600 & 2048x2048 with dynamic English CTC recognition) and real-time object detection (YOLOX-S at 640x640 @ ~9.6 FPS) with system-wide CLI utilities (npu-ocr-pdf, yolox).
  • Vendor NPU & Hardware Acceleration Tools (NPU_tools_vendor): Legacy vendor kernel acceleration stack, hardware monitoring utilities, and acceleration reports for the vendor Allwinner BSP.

About

Builds, binaries, tutorials for OrangePi 4A AllWinner device

Resources

Stars

6 stars

Watchers

2 watching

Forks

Releases

Packages

Contributors

Languages