Skip to content

01GettingStarted.md

Latest commit

 

History

History
216 lines (159 loc) · 8.58 KB

01GettingStarted.md

File metadata and controls

216 lines (159 loc) · 8.58 KB

Artificial Intelligence Applications for ALL

Lecture Notes by idebtor@gmail.com

Getting Started

To get started, do the first thing first:

  1. Read README.
  2. Read Syllabus.
  3. Read '01GettingStarted' - this file
  4. Follow instructions in '01GettingStarted' as soon as possible(ASAP).

This is available at github.com/idebtor/JoyAI.

Join Piazza.

There are two ways to join Piazza, go the www.piazza.com.

  • To join Piazza, you may need the following information and

    • School: Handong Global University
    • Course: Artificial Intelligence Applications for ALL

  • If you have an email address that ends with ~.hgu.edu or ~.handong.edu domain and use it everyday, go the www.piazza.com and follow the instructions in the website.

  • On your request with your email address, I may register it for you. We'll be conducting all class-related discussion here this term. The quicker you begin asking questions on Piazza (rather than via emails), the quicker you'll benefit from the collective knowledge of your classmates and instructors. We encourage you to ask questions when you're struggling to understand a concept—you can even do so anonymously.

Install Anaconda

Anaconda is a Python and R distribution package. It aims to provide everything you need (python wise) for data science "out of the box". It includes:

  • The core python language
  • 200+ python "packages" (libraries)
  • Spyder (IDE/editor) and Jupyter Notebook
  • conda, Anaconda's own package manager, used for updating Anaconda and packages

To install the anaconda:

  • Visit website Anaconda Distribution
    • Choose one of Windows/MacOS/Linux
    • Python 3.x Version Download
  • At the beginning of installation, check the following option
    • Add Anaconda to my PATH environment variable
  • Need help? Follow this guide.

After your installation

Do the following in cmd windows or in PowerShell to check your successful installation; ($ is just a prompt of your console, >>> is a prompt from Python.)

  ```
  $ python
  >>> import tensorflow as tf
  >>> print(tf.__version__)
  1.13.1
  >>> import keras
  Using TensorFlow backend
  ```

Need more installation?

Use the following command if you need more installation of packages (-U for upgrade only):

$ pip install a_package_name
$ pip install -U a_package_name

Install GitHub Desktop

After installation of GitHub Desktop, be a member if already not.

  • Clone the GitHub JoyAI repository into your local computer:

  • How to clone a repository from GitHub:

  • Click 'watch' and 'star' at the top of the web page^^.

  • Then, in your computer, you may have the following github/JoyAI folder as shown below (user may be different in your system.):

    • C:\Users\user\Documents\GitHub\JoyAI

  • Since this JoyAI repository can be updated anytime, keep this local repository as "read-only". Don't code yours here!.

  • Copy them into your own repository or your own local development folders in your computer you can easily access them. They should look like the following:

    ~/JoyAI/dataset
~/JoyAI/deep_class
~/JoyAI/deep_code
~/JoyAI/ppts
~/JoyAI/01GettingStarted.md
~/JoyAI/README.md             
~/JoyAI/Syllabus       
~/JoyAI/Weekly01-02OpenLectures.md

    Note for Multi-screen users: Remove the following file if GitHub Desktop is displayed off-screen. Restart Desktop GitHub. (user below may be different in your system.)

    C:\Users\user\AppData\Roaming\GitHub Desktop\window-state.json

Install Atom.

  1. Atom is a text editor that most professional programmers love nowadays.
  2. Start Atom.
  3. Install some of essential packages recommended for C/C++ programmers listed below:
- Autosave
  - It automatically saves files when the editors loses focus, are destroyed, or when the window is closed. Believe or not, it is disabled by default. __You must check `enabled`__ in config setting or from the Autosave section of the Settings view.

- Markdown-preview
  - Open a rendered version of the Markdown in the current editor with `ctrl-shift-m`.
- File-icons
- Mini-maps

__Themes of my personal preference__:
  - UI Theme: __Atom Dark__
  - Syntax Theme: __Oceanic Next__

__Note for Multi-screen users:__

If Atom is displayed off-screen, do the following:
  1. Alt + Tab to choose the atom window
  2. Alt + Space to open the context menu
  3. Press 'm' to select move
  4. Press any arrow key once
  5. Move your mouse (The misplaced window will follow your cursor.)

Are ready for 'Hello World!' program in Python?

  • Open a console. (You may use cmd or powershell in Windows.)
$ python
>>> print('Hello World!')
>>> 1 + 2
3
>>> exit()
$

A few ways to start Jupyter notebook

Method 1:

  1. Using Anaconda Navigator, choose Jupyter Notebook.

Method 2:

(This option may not work unless you have set PATH environment variable.)

  1. Using File Explorer, navigate to where your Jupyter notebook file is
  2. Using File menu in File Explorer, click Open PowerShell(PS).
  3. At PS console, enter the following:
PS C:\> jupyter notebook

Method 3: This is one-line batch command file that runs Jupyter notebook.

  1. Get a copy of the batch file start_ipynb.bat which is available at https://github.com/idebtor/JoyAI
  2. Place the batch file at the folder where your notebook file is.
  3. Double-click the batch file.

Method 4: (unstable)

Make an association .ipynb extension with Jupyter Notebook.

  1. python -m pip install nbopen
  2. To integrate with your file manager, so you can double click on notebooks to open them, run:
- Linux/BSD: python -m nbopen.install_xdg

- Windows: python -m nbopen.install_win

- Mac: Clone the [repository](https://github.com/takluyver/nbopen) and run ./osx-install.sh

Are ready for cloning another open-source program in GitHub?

Let's clone some open-source packages that uses with MNIST dataset and classifies user's hand-writing digits interactively.

  • Clone the following two github sites
https://github.com/scrambledpie/Drawing-Mnist-and-Cifarizing-image-files
https://github.com/rhammell/mnist-draw

Drawing-Mnist-and-Cifarizing-image-files

Once you clone this source code, you may see the folder named Drawing-Mnist-and-Cifarizing-image-files and ipynb files in your local folder.

The following Jupyter notebook code provides a blank canvas in the notebook so that a user may create their own test samples to input into a trained neural net in Keras. Some code borrowed from Francois Chollet https://github.com/fchollet

  • Run the code cells one by one in DrawMyOwnNumbers.ipynb
  • If you see some error messages, read and attempt to resolve the problems.
  • You may be asked to install more packages/modules. Then how would you do?

Minist-draw (unstable)

  • (I am still experiencing a difficulty to run this program in some machines.)

  • Follow the instructions in README. Then you may be able to start a web server and display a web page that has a user interface getting user's hand-writing digits and recognize them interactively.

  • If you experience an error something like

127.0.0.1 - - [04/Mar/2019 14:55:50] b'ModuleNotFoundError: No module named \'numpy.core._multiarray_umath\'\r\nImportError: numpy.core.multiarray failed to import\r\n\r\nThe above exception was the direct cause of the following exception:\r\n\r\nTraceback (most recent call last):\r\n  File "<frozen importlib._bootstrap>", line 980,
  • install numpy with -U (upgrade) option as shown below: 
    pip install -U numpy

References

What's Next?

  • To learning Python and Numpy from the beginning, visit this site. It is offered by learnpython.org.

One thing I know, I was blind but now I see. John 9:25