機(jī)器學(xué)習(xí)常用Python庫安裝

作者	日期	版本	說明
Dog Tao	2022.06.16	V1.0	開始建立文檔

Anaconda

簡介

Anaconda and Miniconda are distributions of Python and other packages for data science, while Conda is the package manager that installs, updates, and removes them. Anaconda includes hundreds of packages, while Miniconda includes only Conda and its dependencies. Conda can also access different channels, such as the main channel maintained by Anaconda and the conda-forge channel maintained by the package developers. Users can choose between Anaconda Navigator, a graphical user interface, or Conda, a command-line tool, to manage their environments and packages.

Conda官方網(wǎng)站：https://docs.conda.io/en/latest/

Conda is an open source package management system and environment management system that runs on Windows, macOS, and Linux. Conda quickly installs, runs and updates packages and their dependencies. Conda easily creates, saves, loads and switches between environments on your local computer. It was created for Python programs, but it can package and distribute software for any language.

Conda as a package manager helps you find and install packages. If you need a package that requires a different version of Python, you do not need to switch to a different environment manager, because conda is also an environment manager. With just a few commands, you can set up a totally separate environment to run that different version of Python, while continuing to run your usual version of Python in your normal environment.

In its default configuration, conda can install and manage the thousand packages at repo.anaconda.com that are built, reviewed and maintained by Anaconda?.

Conda can be combined with continuous integration systems such as Travis CI and AppVeyor to provide frequent, automated testing of your code.

The conda package and environment manager is included in all versions of Anaconda and Miniconda.

Conda is also included in Anaconda Enterprise, which provides on-site enterprise package and environment management for Python, R, Node.js, Java and other application stacks. Conda is also available on conda-forge, a community channel. You may also get conda on PyPI, but that approach may not be as up to date.

Anaconda官方網(wǎng)站：https://www.anaconda.com/

Anaconda was founded in 2012 by Peter Wang and Travis Oliphant out of the need to bring Python into business data analytics, which was rapidly transforming as a result of emerging technology trends. Additionally, the open-source community lacked an entity that could organize and collectivize it to maximize its impact. Since that time, the Python ecosystem has significantly expanded, with Python being the most popular programming language used today. Alongside this expansion, Anaconda has provided value to students learning Python and data science, individual practitioners, small teams, and enterprise businesses. We aim to meet every user where they are in their data science journey. Anaconda now has over 300 full-time employees based in the United States, Canada, Germany, United Kingdom, Australia, India, and Japan. We are proud to serve over 35 million users worldwide.

使用

參考文檔：Anaconda conda常用命令：從入門到精通

在anaconda官網(wǎng)搜索包：https://anaconda.org/

鏡像源配置

參考文檔：conda操作之更新源和刪除源

• 查看鏡像源

conda config --show channels

• 永久添加鏡像源

使用conda config --add channels URL命令，以添加清華源為例：

conda config --add channels https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/free/
conda config --add channels https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/main/
conda config --add channels https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud/conda-forge/

• 移除鏡像源

使用conda config --remove channels URL命令，以移除清華源為例：

conda config --remove channels https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/main
conda config --remove channels https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/free
conda config --remove channels https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud/conda-forge/

• 設(shè)置搜索時(shí)顯示通道地址

conda config --set show_channel_urls yes

• 臨時(shí)指定使用某個(gè)鏡像源下載

使用conda的參數(shù)-c指定鏡像源的地址，例如想在清華鏡像源下載opencv包：

conda install opencv -c https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/main/

國內(nèi)鏡像源舉例：

1. 清華源

conda config --add channels  https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/main
conda config --add channels  https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/free
conda config --add channels  https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud/conda-forge/

2. 中科大源

conda config --add channels https://mirrors.ustc.edu.cn/anaconda/pkgs/main/
conda config --add channels https://mirrors.ustc.edu.cn/anaconda/pkgs/free/
conda config --add channels https://mirrors.ustc.edu.cn/anaconda/cloud/conda-forge/

3. 北京外國語大學(xué)源

conda config --add channels  https://mirrors.bfsu.edu.cn/anaconda/pkgs/main
conda config --add channels  https://mirrors.bfsu.edu.cn/anaconda/pkgs/free
conda config --add channels  https://mirrors.bfsu.edu.cn/anaconda/cloud/conda-forge/

4. 上海交大源

conda config --add channels  https://mirrors.sjtug.sjtu.edu.cn/anaconda/pkgs/main/
conda config --add channels  https://mirrors.sjtug.sjtu.edu.cn/anaconda/pkgs/free/
conda config --add channels  https://mirrors.sjtug.sjtu.edu.cn/anaconda/cloud/conda-forge/

5. 豆瓣源

conda config --add channels https://pypi.doubanio.com/simple/

Pip

簡介

官網(wǎng)：https://pypi.org/project/pip/

pip is the package installer for Python. You can use pip to install packages from the Python Package Index and other indexes.

鏡像源配置

參考文檔：pip國內(nèi)鏡像源配置

pip官方軟件源 https://pypi.python.org/simple

國內(nèi)鏡像源舉例：

1. 阿里云 https://mirrors.aliyun.com/pypi/simple/

2. 中國科技大學(xué) https://pypi.mirrors.ustc.edu.cn/simple/

3. 豆瓣 https://pypi.douban.com/simple

4. 中國科學(xué)院 https://pypi.mirrors.opencas.cn/simple/

5. 清華大學(xué) https://pypi.tuna.tsinghua.edu.cn/simple/

• 臨時(shí)指定使用某個(gè)鏡像源下載

使用pip的參數(shù)-i指定鏡像源的地址，例如想在阿里云鏡像源下載Pillow包

pip install -i https://mirrors.aliyun.com/pypi/simple Pillow

CUDA

• 顯卡型號支持檢查：https://developer.nvidia.com/cuda-gpus

• Archived ReleasesCUDA Toolkit下載：https://developer.nvidia.com/cuda-toolkit-archive

• 技術(shù)教程：https://blog.csdn.net/Mind_programmonkey/article/details/99688839

Pytorch

官方安裝說明：https://pytorch.org/get-started/locally/

安裝舊版本

Installing previous versions of PyTorch： https://pytorch.org/get-started/previous-versions/

以適配CUDA 11.3的版本為例：

conda install pytorch==1.12.1 torchvision==0.13.1 torchaudio==0.12.1 cudatoolkit=11.3 -c pytorch

TensorFlow

官方安裝說明：https://tensorflow.google.cn/install?hl=zh-cn

GPU支持說明

官方安裝說明：https://tensorflow.google.cn/install/gpu?hl=zh-cn

DGL

簡介

官網(wǎng)：https://www.dgl.ai/

In the last few years, deep learning has enjoyed plenty of extraordinary successes. Many challenging tasks have been solved or close to being solved by Deep Learning, such as image recognition, rich-resource machine translation, game playing. These were made possible by a set of techniques that are composed of a number of representationally powerful building-blocks, such as convolution, attention and recurrence, applied to images, video, text, speech and beyond.The development and deployment of these techniques often depend on the simple correlation of the given data; for example, CNN is based on the spatial correlation between nearby pixels while RNN family dwells on the assumption that its input is sequence-like.More recently, there has been a steady flow of new deep learning research focusing on graph-structured data. Some of them are more conventional graph related problems, like social networks, chemical molecules and recommender systems, where how the entity interacts with its neighborhood is as informative as, if not more than, the features of the entity itself.Some others nevertheless have applied graph neural networks to images, text or games. Very broadly speaking, any of the data structures we have covered so far can be formalized to graphs. For instance an image can be seen as grid of pixel, text a sequence of words… Together with matured recognition modules, graph can also be defined at higher abstraction level for these data: scene graphs of images or dependency trees of language.To this end, we made DGL. We are keen to bringing graphs closer to deep learning researchers. We want to make it easy to implement graph neural networks model family. We also want to make the combination of graph based modules and tensor based modules (PyTorch or MXNet) as smooth as possible.

安裝

官方安裝說明：https://www.dgl.ai/pages/start.html

以適配CUDA 11.3的版本為例：

# If you have installed dgl-cudaXX.X package, please uninstall it first.
conda install -c dglteam/label/cu113 dgl

DGLLife

DGL-LifeSci官網(wǎng)：https://lifesci.dgl.ai/index.html

DGL-LifeSci is a python package for applying graph neural networks to various tasks in chemistry and biology, on top of PyTorch, DGL, and RDKit. It covers various applications, including:

• Molecular property prediction
• Generative models
• Reaction prediction
• Protein-ligand binding affinity prediction

DGL-LifeSci is free software; you can redistribute it and/or modify it under the terms of the Apache License 2.0. We welcome contributions. Join us on GitHub.

• 在anaconda官網(wǎng)搜索包：https://anaconda.org/

conda install -c conda-forge dgllife

RDKit

官網(wǎng)：https://rdkit.org/

RDKit documentation：https://rdkit.org/docs/index.html

conda install -c conda-forge rdkit

pip install rdkit

scikit-multilearn

官網(wǎng)：http://scikit.ml/

文檔：http://scikit.ml/api/skmultilearn.html

源碼：https://github.com/scikit-multilearn/scikit-multilearn

pip install scikit-multilearn