tests

Miscellaneous AI-related materials in Python

GitHub code, https://github.com/jinghuazhao/tests/blob/main/AI/python/

Command history is ~/.python_history for interactive Python sessions, while use of Python virtural environment is shown in MDL.sh¹.

• AI and ML for Coders in PyTorch, https://github.com/lmoroney/PyTorch-Book-FIles
• Build a Text-to-Image Generator (from Scratch): With transformers and diffusions, https://github.com/markhliu/txt2img
• Deep Learning Crash Course, https://github.com/DeepTrackAI/DeepLearningCrashCourse
• Deep Learning with Python, 3e, https://github.com/fchollet/deep-learning-with-python-notebooks
• Generative AI with Python and PyTorch, 2e, https://github.com/PacktPublishing/Generative-AI-with-Python-and-PyTorch-Second-Edition
• Graph Neural Network in Action, https://github.com/keitabroadwater/gnns_in_action
• Math and Architectures of Deep Learning, https://github.com/krishnonwork/mathematical-methods-in-deep-learning-ipython
• Math for Deep Learning, https://github.com/rkneusel9/MathForDeepLearning
• Practical Deep Learning, 2e, https://github.com/rkneusel9/PracticalDeepLearning2E
• Ultimate Neural Network Programming with Python², https://github.com/OrangeAVA/Ultimate-Neural-Network-Programming-with-Python

.ipynb

Code extraction is furnished in two steps as follows,

# --> markdown
source rds/software/py3.11/bin/activate
jupyter nbconvert --to markdown Chapter8.ipynb
# --> Python
module load ceuadmin/node/
npm install -g codedown
which codedown
cat Chapter8.md | codedown python > Chapter8.py

Visual Studio code generally does a very good job but it is greatly faciliated by Google Colab, https://colab.research.google.com/. Notably, follow “Runtime” –> “Change runtime type” to enable GPU and/or “Copy path” from the folder listing. In the case of Image_segmentation_pipeline.ipynb, it is also handy to invoke command palette via Ctrl-Shift-P and enter terminal.

One can also post the GitHub address to https://nbviewer.org, e.g., by replacing stable_diffusion.ipynb below with a .ipynb in the folder, e.g., https://nbviewer.org/github/jinghuazhao/tests/blob/main/AI/python/stable_diffusion.ipynb.

Google colab

It is handy to clone a GitHub repository into Google drive and use it from there, e.g., insert these lines

from google.colab import drive
import os

drive.mount('/content/drive')

BASE = '/content/drive/MyDrive'
REPO = f'{BASE}/DeepLearningCrashCourse'

if not os.path.isdir(REPO):
    %cd $BASE
    !git clone https://github.com/DeepTrackAI/DeepLearningCrashCourse

%cd $REPO
!ls -R
%cd Ch01_DNN_classification/ec01_1_neuron_class_1d

into neuro_class_1d.ipynb then the notebook can run smoothly.

---

1. virtual environment

   #!/usr/bin/bash
   
   # Math for Deep Learning, Chapter 10, <https://github.com/rkneusel9/MathForDeepLearning>
   
   module load python/3.9.12/gcc/pdcqf4o5
   python -m venv venv
   source venv/bin/activate
   pip install opencv-python
   pip install scikit-learn
   pip install keras
   pip install tensorflow
   python < iris.py
   python < nn_by_hand.py
   python < mnist.py
   
   # Under VSCode, the virtual environment is also visible once PATH/PYTHONPATH is added.
   # notedown ResNet18.md > ResNet18.ipynb
   

   ↩

2. cambridge-ceu fork

   There are a number of changes in the forked repository, https://github.com/cambridge-ceu/Ultimate-Neural-Network-Programming-with-Python,

   • model/att_net.py is actually model/att_unet.py
   • changes of file paths
   • rename Chatper –> Chapter for 6-8, 10.

   which have now been merged into the master branch. ↩