RK3399pro debian10 系统下安装RKNN-Toolkit 1.7.0

RK3399pro debian10 系统下安装rknn-toolkit 1.7.0

本次安装方法是基于RK3399pro 官方开发板（RK_EVB_RK3399PRO_LP3S178P332SD8_V14_20191108_LINT)

1. 首先要安装镜像，将镜像root分区3GB变大，

方法如下：下载镜像链接：链接：百度网盘 请输入提取码

提取码：ubls

root@linaro-alip:/#

root@linaro-alip:/# fdisk -l

Disk /dev/ram0: 4 MiB, 4194304 bytes, 8192 sectors

Units: sectors of 1 * 512 = 512 bytes

Sector size (logical/physical): 512 bytes / 4096 bytes

I/O size (minimum/optimal): 4096 bytes / 4096 bytes

Disk /dev/mmcblk0: 14.6 GiB, 15634268160 bytes, 30535680 sectors

Units: sectors of 1 * 512 = 512 bytes

Sector size (logical/physical): 512 bytes / 512 bytes

I/O size (minimum/optimal): 512 bytes / 512 bytes

Disklabel type: gpt

Disk identifier: F90D0000-0000-4A02-8000-41A5000019ED

Device           Start      End  Sectors  Size Type

/dev/mmcblk0p1   16384    24575     8192    4M unknown

/dev/mmcblk0p2   24576    32767     8192    4M unknown

/dev/mmcblk0p3   32768    40959     8192    4M unknown

/dev/mmcblk0p4   40960   106495    65536   32M unknown

/dev/mmcblk0p5  106496   303103   196608   96M unknown

/dev/mmcblk0p6  303104   368639    65536   32M unknown

/dev/mmcblk0p7  368640   499711   131072   64M unknown

/dev/mmcblk0p8  499712  1548287  1048576  512M unknown

/dev/mmcblk0p9 1548288 30535646 28987359 13.8G unknown

Disk /dev/mmcblk0boot1: 4 MiB, 4194304 bytes, 8192 sectors

Units: sectors of 1 * 512 = 512 bytes

Sector size (logical/physical): 512 bytes / 512 bytes

I/O size (minimum/optimal): 512 bytes / 512 bytes

Disk /dev/mmcblk0boot0: 4 MiB, 4194304 bytes, 8192 sectors

Units: sectors of 1 * 512 = 512 bytes

Sector size (logical/physical): 512 bytes / 512 bytes

I/O size (minimum/optimal): 512 bytes / 512 bytes

root@linaro-alip:/# resize2fs /dev/mmcblk0p9

resize2fs 1.44.5 (15-Dec-2018)

Filesystem at /dev/mmcblk0p9 is mounted on /; on-line resizing required

old_desc_blocks = 1, new_desc_blocks = 2

[  202.156834] EXT4-fs (mmcblk0p9): resizing filesystem from 912107 to 3623419 blocks

[  202.372105] EXT4-fs (mmcblk0p9): resized filesystem to 3623419

The filesystem on /dev/mmcblk0p9 is now 3623419 (4k) blocks long.

root@linaro-alip:/# [  223.159444] wl_run_escan: LEGACY_SCAN sync ID: 6, bssidx: 0

root@linaro-alip:/#

root@linaro-alip:/# ls

bin   dev  home  lost+found  media  opt   root  sbin  sys     tmp  var

boot  etc  lib   md5sum.txt  mnt    proc  run   srv   system  usr

root@linaro-alip:/# df -

df: -: No such file or directory

root@linaro-alip:/# df -h

Filesystem      Size  Used Avail Use% Mounted on

/dev/root        14G  2.4G   11G  18% /

devtmpfs        1.9G  8.0K  1.9G   1% /dev

tmpfs           1.9G     0  1.9G   0% /dev/shm

tmpfs           1.9G   17M  1.9G   1% /run

tmpfs           5.0M  4.0K  5.0M   1% /run/lock

tmpfs           1.9G     0  1.9G   0% /sys/fs/cgroup

tmpfs           379M  8.0K  379M   1% /run/user/1000

/dev/mmcblk0p8  504M  2.3M  502M   1% /media/linaro/a4a71807-5b42-4064-a1e4-4faaa29191ca

/dev/mmcblk0p7   63M   13M   48M  21% /media/linaro/b3115340-7e34-4fd5-894e-5766b617b5b9

tmpfs           379M     0  379M   0% /run/user/0

1. 安装必要的软件：

由于rknn-toolkit 中tensorflow 依赖python 3.6的版本，而SDK默认提供的是Python 3.7 的版本，默认使用pip 进行安装会提示如下错误：

(venv) root@linaro-alip:/venv# ls

venv

(venv) root@linaro-alip:/venv# pip install --upgrade tensorflow

[24666.992199] wl_run_escan: LEGACY_SCAN sync ID: 394, bssidx: 0

ERROR: Could not find a version that satisfies the requirement tensorflow (from versions: none)

ERROR: No matching distribution found for tensorflow

(venv) root@linaro-alip:/venv#

需要单独下载arm tensorflow 进行安装。

报错解决方法：

下载tensorflow-1.15.5-cp37-cp37m-linux_aarch64.whl， 用adb push到板端，执行 pip install tensorflow-1.15.5-cp37-cp37m-linux_aarch64.whl 

下载链接：https://tf.kmtea.eu/whl/stable.html 或者百度网盘：

2.1  

sudo apt-get update

sudo apt-get install cmake gcc g++ libprotobuf-dev protobuf-compiler

sudo apt-get install liblapack-dev libjpeg-dev zlib1g-dev

sudo apt-get install python3-dev python3-pip python3-scipy

执行以下命令更新 pip

pip3 install --upgrade pip

安装 Python 打包工具

pip3 install wheel setuptools

安装依赖包 h5py/gluoncv

sudo apt-get build-dep python3-h5py && \ 

pip3 install h5py 

pip3 install gluoncv

安装 TensorFlow

pip install /rknn-toolkit/packages/tensorflow-1.15.5-cp37-cp37m-linux_aarch64.whl

安装 torch 和 torchvision

暂时先用不着，先不安装

安装 opencv-python

pip3 install opencv-python==4.3.0.38 -i Simple Index --default-timeout=200

安装 RKNN-Toolkit

首先：将rknn-toolkit github demo 下载下来：

git clone https://github.com/rockchip-linux/rknn-toolkit

然后：下载rknn-toolkit 1.7.1  adb push 到板端 

下载链接：https://github.com/rockchip-linux/rknn-toolkit 或者百度网盘：链接：百度网盘 请输入提取码提取码：6jp9