Quick Start

[中文]

Before you start

  • BK AI vs SMP: This repository contains solution and application code. Firmware builds depend on BK AVDK SMP via SDK_DIR. Drivers, RTOS, Wi-Fi/BLE, etc. are provided by SMP. See Armino AI Solution Introduction.

  • Choosing a project: For beken_genie, volc_rtc, ai_camera, see Choosing a project in Reference Projects.

Note

If you cannot access GitLab (enterprise-only), use the GitHub clone commands below or contact your Beken representative.

1. Environment Setup

1.1 Download Armino SMP SDK:

GitLab: You can download the Armino SMP code from GitLab:

mkdir -p ~/armino
cd ~/armino
git clone https://gitlab.bekencorp.com/armino/bk_avdk_smp.git -b release/v3.1.1

GitHub: You can also clone from bekencorp/bk_avdk_smp on GitHub:

mkdir -p ~/armino
cd ~/armino
git clone https://github.com/bekencorp/bk_avdk_smp.git -b release/v3.1.1

1.2 Download AI Solution Code:

GitLab: You can download the Armino AI Solution code from GitLab:

mkdir -p ~/armino
cd ~/armino
git clone https://gitlab.bekencorp.com/armino/smp_solution/bk_solution_ai.git -b release/v3.1.1

GitHub: You can also clone from bekencorp/bk_solution_ai on GitHub:

mkdir -p ~/armino
cd ~/armino
git clone https://github.com/bekencorp/bk_solution_ai.git -b release/v3.1.1

Note

On Windows, run git config --global core.autocrlf false before cloning to avoid line-ending issues that break the build.

1.3 Environment Deployment and Build

We provide two deployment options: a local build environment and a Docker-based build environment. You may choose either based on your hardware and habits; we generally recommend the local build environment.

The local deployment option currently supports building on Windows and Linux only.

Docker deployment uses containers for environment setup and builds. It supports efficient compilation on Linux, macOS, and Windows. With Docker, you do not need to manually install the libraries and toolchains required for building, which greatly simplifies setup and compilation. This option suits users who are familiar with Docker. If you are not, or your network prevents using Docker, use the local deployment option instead.

2. Build Project

The examples below use beken_genie. For volc_rtc, replace beken_genie in the path with volc_rtc; output is under build/bk7258/<project>/package.

Method 1: Direct Build

cd ~/armino/bk_solution_ai/projects/beken_genie
make clean SDK_DIR=~/armino/bk_avdk_smp
make bk7258 SDK_DIR=~/armino/bk_avdk_smp

Method 2: Specify SDK Path Using Export

cd ~/armino/bk_solution_ai/projects/beken_genie
export SDK_DIR=~/armino/bk_avdk_smp
make clean
make bk7258

Method 3: Using Docker (Linux/Mac)

cd ~/armino/bk_solution_ai/projects/beken_genie
export SDK_DIR=~/armino/bk_avdk_smp
./dbuild.sh make clean
./dbuild.sh make bk7258

Method 4: Using Docker (Windows PowerShell)

cd C:\armino\bk_solution_ai\projects\beken_genie
$env:SDK_DIR = "C:\armino\bk_avdk_smp"
.\dbuild.ps1 make clean
.\dbuild.ps1 make bk7258

3. Flash Firmware to Device

After compilation, all-app.bin is generated under build/bk7258/beken_genie/package (replace beken_genie with your project folder name if different). Flash it to the board with the Beken tool.

3.1 Resource File Flashing

4. APP Registration and Download

5. FAQ

  • Where should ``SDK_DIR`` point? To the root of the cloned BK AVDK SMP tree (where Makefile and tools live), not the BK AI repository root.

  • Missing toolchain or script errors? Complete SMP setup in Local Deployment or Docker Deployment first; for Docker, ensure Docker is installed and running.

  • Slow Docker image pull? Use the mirror / download options described in Docker Deployment.