ivan-bilan / Deep-Learning-for-NLP-Tutorials

Tutorials for using Tensorflow and Keras for NLP

Installation Guide

Use Python 3.5 or higher

Step 1, clone the GitHub project

Go to https://github.com/ivan-bilan/Tensorflow-for-NLP-Tutorials Press: "Clone and Download"

Navigate to the folder you want to install the project in and do:

git clone https://github.com/ivan-bilan/Tensorflow-for-NLP-Tutorials.git

Step 2, Jupyter Notebooks

Install Jupyter Notebook

pip3 install jupyter

Starting the Notebook

In PyCharm

If you use PyCharm, simply use the Start button on the main project file tensorflow_basics.ipynb and copy the generated URL with the security token, e.g.: http://127.0.0.1:8888/?token=11cff7e3126674b2cdd8093a1xxx954a67. Open the URL in your browser. That's it!

In terminal

Run from terminal:

jupyter notebook

it will navigate to your home folder, go to the project folder using the Jupyter file browser.

For both solutions, use Chrome or Chromium. Firefox won't work with certain examples.

Step 3, Tensorflow and related packages

Now let's install tensorflow and tflearn.

On your own laptop

If you have your own laptop with you with sudo access, just run there commands in your terminal:

pip3 install tensorflow 
pip3 install tflearn
pip3 install numpy
pip3 install pandas==0.19.2

University PC

Option 1

Install pip packages only for your local user:

pip3 install --user tensorflow 
pip3 install --user tflearn
pip3 install --user numpy
pip3 install --user pandas==0.19.2

Option 2

If you have PyCharm installed, just do the following:

File > Project xx > Project Interpreter > Select the "+" on the right top side > search for ternsorflow > press "Install Package"

Should work without any errors. Do the same for tflearn, numpy and pandas.

Alternative: Running Jupyter Notebook in the cloud

If you want to skip local installation, use cocalc.com sign-up with GitHub/Google

1. create new project
2. choose JupyterNotebook as type
3. go to "Kernel" menu, and choose Python 3 (Anaconda)

Now you have an IpythonPython Notebook with tensorflow installed and running in the cloud.