AI工具安装记录-install

1.安装

1.1 Anaconda

进入终端：
conda deactivate，退出anaconda环境。
conda activate， 激活环境。

conda create -n tf2 python=3.6

1.2 tensorflow安装

ubuntu16.04+anaconda3+python3.6

conda create -n tf python=3.6
python -m pip install tensorflow-cpu -i https://pypi.douban.com/simple --user


pip install --ignore-installed --upgrade https://mirrors.tuna.tsinghua.edu.cn/tensorflow/linux/cpu/tensorflow-1.4.0-cp36-cp36m-linux_x86_64.whl 

sudo pip install jupyter 

sudo chmod -R 777 Anaconda

source activate tf

忽略tf警告:

import os
os.environ['TF_CPP_MIN_LOG_LEVEL'] = '2' 

1.3 opencv

pip3 install --user opencv-python==4.2.0.34(最新版容易死机)

pip show opencv-python

Location: /home/hjz/.local/lib/python3.6/site-packages

Location: /home/hjz/anaconda3/envs/tf/lib/python3.6/site-packages

pip --default-timeout=100 install opencv-contrib-python==4.2.0.34 -i https://pypi.douban.com/simple

pip3 --default-timeout=100 install --user opencv-python==4.2.0.34 -i https://pypi.douban.com/simple

1.4 face_recognition

1.4.1环境搭建

安装 git 、cmake 、 python-pip

安装 git

$ sudo apt-get install -y git

安装 cmake

$ sudo apt-get install -y cmake

安装 python-pip

$ sudo apt-get install -y python-pip

1.4.2安装编译 dlib

编译dlib前先安装 boost

$ sudo apt-get install libboost-all-dev

开始编译dlib

克隆dlib源代码

$ git clone https://github.com/davisking/dlib.git
$ cd dlib
$ mkdir build
$ cd build
$ cmake … -DDLIB_USE_CUDA=0 -DUSE_AVX_INSTRUCTIONS=1
$ cmake --build .（注意中间有个空格）
$ cd …
$ sudo python setup.py install --yes USE_AVX_INSTRUCTIONS --no DLIB_USE_CUDA

1.4.3安装 face_recognition

安装 face_recognition

$ sudo pip install face_recognition

1.5 jupyter kernel error

base 安装 nb_conda

envs 安装 ipykernel

jupyter 启动：base下jupyter notebook +文件夹路径

1.6 pip --default-timeout=100 install

pip --default-timeout=100 install pandas

pip --default-timeout=100 install sklearn

1,7 sougou 输入法乱码

killall fcitx

1.8 MXnet-gluncv

pip --default-timeout=100 install gluoncv --upgrade

pip --default-timeout=100 install mxnet-mkl --upgrade

1.9 pytorch

torch官网cpu,pip --default-timeout=100 install +.whl

1.10 meshroom

pip --default-timeout=100 install -r requirements.txt

LD_LIBRARY_PATH=/usr/lib/nvidia PYTHONPATH=$PWD python meshroom/ui 不行

下载安装包-更改权限sudo chmod 777 Meshroom打开—>./Meshroom

1.11 打印进度条

from time import sleep
from tqdm import tqdm

for i in tqdm(range(1,500)):
    sleep(0.01)
sleep(0.5)

1.12 yolov4

cd darknet
make

./darknet detect cfg/yolov4.cfg yolov4.weights data/dog.jpg

2.work

2.1深度学习可视化

2.1.1 可视化工具

1.https://poloclub.github.io/cnn-explainer/

CNN在线交互可视化工具

2,卷积https://cs231n.github.io/assets/conv-demo/index.html

2.2. 文字转语音

https://zhuanlan.zhihu.com/p/75819181

https://www.ffkuaidu.com/