Win10 drqa installation

Before the installation DrQA Meet some pits , stay windows System ,cenos All installed , It didn't work , Last in Ubuntu To run successfully , I've sorted it out here Ubuntu Installation steps on the system

Vmware install Ubuntu

Vmware The installation steps are simple, so the user doesn't have to say . Then install Ubuntu Here is my version Ubuntu download
I have a desktop version installed here , It is recommended that the memory be as large as possible when installing the system , I run memory with 8G, For storage space 40G, Because the data will be downloaded later 7.5G

Installation of required tools and software

1. Git install

sudo apt-get install git

According to the prompt
apt-get update

An error has been encountered here （ If not, please ignore ）
ubuntu Common mistakes –Could not get lock /var/lib/dpkg/lock solve
Install the program through the terminal sudo apt-get install xxx An error occurred when ：
E: Could not get lock /var/lib/dpkg/lock - open (11: Resource temporarily unavailable)
E: Unable to lock the administration directory (/var/lib/dpkg/), is another process using it
The problem may be that another program is running , Causes resources to be locked unavailable . The reason for resource lock may be that the installation or update did not complete normally last time , And then there is this situation , The solution is simple ：

Type the following two sentences into the terminal

sudo rm /var/cache/apt/archives/lock

sudo rm /var/lib/dpkg/lock

And then again sudo apt-get install git
You can make it

2. Anaconda install

Anaconda It's a scientific calculation Python Distribution version , Support Linux, Mac, Windows System , It provides package management and environment management functions , It's easy to solve multiple versions python Coexist 、 Switching and various third-party package installation problems .Anaconda utilize an instrument / command conda To carry out package and environment Management of , And it already contains Python And related tools . The installation steps are as follows ：
（1） from https://repo.continuum.io/archive/index.html Download the corresponding version of Anaconda edition , Unzip to the specified location and enter the directory .
（2） Run the command bash Anaconda3-5.2.0-Linux-x86_64.sh Installation .
（3） function echo 'export PATH="~/anaconda2/bin:$PATH"' >> ~/.bashrc To configure .
（4） Run the command source ~/.bashrc Make configuration effective .
Command input anaconda -V Verify successful installation

And then the problem （ If not, please ignore ）
Ubuntu Next Anaconda3 appear conda：command not found（ Command not found ）
problem ：anaconda: command not found
Solution ：
open Terminal
1、 Use command ：sudo apt install vim install vim Text editor
2、 Use command ：vim ~/.bashrc Modify environment variables
3、 Add a command at the end of the text ：export PATH=~/anaconda3/bin:$PATH
4、 Restart environment variables ：source ~/.bashrc
5、 Input ：anaconda -V
6、 Show ：anaconda Command line client (version 1.6.14)

3. PyTorch install

PyTorch One is based on Python Scientific computing package for , yes DrQA Required computing package for . Be careful ！DrQA It is described on the official website ：DrQA requires Linux/OSX and Python 3.5 or higher. It also requires installing PyTorch (version 0.4.0 is not supported yet), So it is recommended to install 0.4 Following PyTorch, You can visit https://pytorch.org/ Generate the installation command according to the corresponding configuration .
　　 It's loaded with Anaconda after ,

conda install pytorch -c pytorch
ip3 install torchvision

Can be installed automatically PyTorch And dependency packages .

4. DrQA install

This part is in DrQA Of github There is a very detailed introduction on . The specific process is as follows ：

（1） install DrQA And dependency packages

git clone https://github.com/facebookresearch/DrQA.git 
cd DrQA
pip install -r requirements.txt 
python setup.py develop

Requirements.txt Recorded DrQA Various packages on which the system depends , After executing the command, it will automatically download, install and update .

Report errors
unable to execute ‘gcc’: No such file or directory
error: Setup script exited with error: command ‘gcc’ failed with exit status 1

resolvent ：
sudo apt-get install python-dev

（2） install tokenizer

tokenizer There are two options , One is Stanford CoreNLP, The other is Spacy. The description on the official website is as follows ：
　　If you use the CoreNLPTokenizer or SpacyTokenizer you also need to download the Stanford CoreNLP jars and spaCy enmodel, respectively. If you use Stanford CoreNLP, have the jars in your java CLASSPATH environment variable, or set the path programmatically with:
　　import drqa.tokenizers
　　drqa.tokenizers.set_default(‘corenlp_classpath’,’/your/corenlp/classpath/*’)

DrQA default tokenizer by CoreNlp, stay DrQA Under the folder, you can run commands ./install_corenlp.sh Download and install .
After installation, you can choose to add it to the environment variable automatically , But sometimes DrQA Adding to environment variables automatically is incorrect , It is recommended that you add it manually according to the official website . Use... If necessary spacy Can run commands python -m spacy download en.

（3） Download datasets , Word vectors, etc

Run the command to automatically download data sets and word vectors bash ./download.sh There is a large amount of downloads , about 7.5G.
　　 thus DrQA installation is complete , function python scripts/pipeline/interactive.py To try DrQA Of demo.

Reference link ：
DrQA Official website
DrQA Installation manual