将M.2 Coral安装到Raspberry Pi 5

介绍

Coral M.2 Accelerator 双Edge TPU是一个M.2模块，为具有可用M.2 E-key插槽的现有系统和产品带来两个Edge TPU协处理器。

Raspberry Pi第五代旗舰开发计算机配备强大的2.4GHz 64位四核Arm处理器和800MHz VideoCore VII GPU，提供令人印象深刻的图形性能。它提供先进的摄像头支持、多功能连接和增强的外设，非常适合多媒体、游戏和工业任务。

本wiki将向您展示如何将Coral M.2 Accelerator安装到Raspberry Pi 5，最后我们将测试Coral M.2 Accelerator。

准备硬件

Raspberry Pi 5 8GB	Raspberry Pi M.2 HAT+	Coral M.2 Accelerator B+M key
立即获取 🖱️	立即获取 🖱️	立即获取 🖱️

安装硬件

安装Python3.8

Coral软件仅支持Python3.6-Python3.9，但最新版本的Raspberry Pi OS Python是Python3.11。所以我们需要安装Python3.8。

步骤1：更新系统

打开终端并运行以下命令来更新系统：

sudo apt update
sudo apt full-upgrade

Step 2: Install requirements

Open a terminal and run the following commands to install the required packages:

sudo apt-get install -y build-essential tk-dev libncurses5-dev libncursesw5-dev libreadline6-dev libdb5.3-dev libgdbm-dev libsqlite3-dev libssl-dev libbz2-dev libexpat1-dev liblzma-dev zlib1g-dev libffi-dev tar wget vim

Step 3: Download Python3.8

Open a terminal and run the following commands to download Python3.8:

wget https://www.python.org/ftp/python/3.8.0/Python-3.8.0.tgz

Step 4: Install Python3.8

Open a terminal and run the following commands to install Python3.8:

sudo tar zxf Python-3.8.0.tgz
cd Python-3.8.0
sudo ./configure --enable-optimizations
sudo make -j 4
sudo make altinstall
cd ..

Step 5: Check Python3.8

Open a terminal and run the following commands to check Python3.8:

python3.8 -V

结果应该为：

步骤 6：使用 Python3.8 创建虚拟环境

python3.8 -m venv coral_venv

Configure Hardware Settings

Open a terminal and run the following commands to open config.txt:

sudo nano /boot/firmware/config.txt

And then add the following text to config.txt:

[all]
# Enable the PCIe External connector.
dtparam=pciex1
kernel=kernel8.img
# Enable Pineboards Hat Ai
dtoverlay=pineboards-hat-ai

Save and close the file by pressing CTRL+X, then Y to confirm. And then reboot the system.

sudo reboot

检查内核：

打开终端并运行以下命令来检查内核：

note

确保您的内核版本为 6.6.30 或更高版本

uname -a

Install the PCIe Driver and Edge TPU Runtime

Step 1: Enter the virtual environment

source coral_venv/bin/activate

Step 2: Install Edge TPU Runtime

Add the Google Coral Edge TPU package repository

echo "deb https://packages.cloud.google.com/apt coral-edgetpu-stable main" | sudo tee /etc/apt/sources.list.d/coral-edgetpu.list

curl https://packages.cloud.google.com/apt/doc/apt-key.gpg | sudo apt-key add -

sudo apt-get update

Install necessary packages and Edge TPU Runtime

sudo apt-get install cmake libedgetpu1-std devscripts debhelper dkms dh-dkms

Step 3: Install the PCIe driver

git clone https://github.com/google/gasket-driver.git
cd gasket-driver
sudo debuild -us -uc -tc -b
cd ..
sudo dpkg -i gasket-dkms_1.0-18_all.deb

设置 udev 规则 添加 udev 规则来管理设备权限

打开终端并运行以下命令来设置 udev 规则：

sudo sh -c "echo 'SUBSYSTEM==\"apex\", MODE=\"0660\", GROUP=\"apex\"' >> /etc/udev/rules.d/65-apex.rules"

sudo groupadd apex

sudo adduser $USER apex

sudo reboot

Check Edge TPU

lspci -nn | grep 089a

结果应该为：

ls /dev/apex_0

安装 PyCoral 库并测试 Edge TPU

步骤 1：安装 PyCoral 库

source coral_venv/bin/activate
pip install --upgrade pip
python3 -m pip install --extra-index-url https://google-coral.github.io/py-repo/ pycoral~=2.0

Step 2: Test the Edge TPU

Install resources for the example:

mkdir coral && cd coral
git clone https://github.com/google-coral/pycoral.git
cd pycoral
bash examples/install_requirements.sh classify_image.py

Test the Edge TPU:

python3 examples/classify_image.py \
--model test_data/mobilenet_v2_1.0_224_inat_bird_quant_edgetpu.tflite \
--labels test_data/inat_bird_labels.txt \
--input test_data/parrot.jpg

结果应该是：

结果

我们已经成功在树莓派 5 上安装了 M.2 Coral 加速器并测试了 Edge TPU。我们还在 Coral M.2 加速器上运行了 YOLOv8s 模型，使用 int8 量化，输入尺寸为 640x640，批处理大小为 1。推理时间约为 800-1000ms，相当于大约 1.1 帧每秒（FPS）。

技术支持与产品讨论

感谢您选择我们的产品！我们在这里为您提供不同的支持，以确保您使用我们产品的体验尽可能顺畅。我们提供多种沟通渠道，以满足不同的偏好和需求。