/kaikai_AI

Learning ML

Primary LanguagePythonMIT LicenseMIT

AI Learning with kaikai

Start Developing

You have to install python3, and module written in ./requirements.txt.

To run cnn with THE MNIST DATABASE of handwritten digits, please download the data set from here and locate the file to dataset/input/.

Create out folder to save graph, and numpy files.

Format

Install yapf (Yet Another Python Formatter).

$ pip3 install yapf

Our format settings is here
if you use Visual Studio Code, please set as below.

.vscode/settings.json

{
  "python.formatting.provider": "yapf",
  "python.formatting.yapfArgs": [
    "--style",
    "{indent_width: 4,ARITHMETIC_PRECEDENCE_INDICATION:True,COLUMN_LIMIT:120,BLANK_LINES_AROUND_TOP_LEVEL_DEFINITION=1,based_on_style: google}"
  ]
}

Tests

We use pytest to test our code.
Please install and run.
The tests run with python3.8.

$ pip3 install pytest
$ pytest -v

We add tests when we confirm that the theory is correct.

Comments

We write comments actively, because we can see more easily.
Comments are written English or Japanese sentence, and required in situations like below.

  • Attention needed
  • Multiple tasks in one function
  • Complex tasks

Road Map & Status

No. Road Map Status
1 DNN 🐥
2 CNN 🐣
3 RNN 🥚
4 RL(Q-learning) 🥚
4 Auto Encoder 🥚
5 SVM 🥚

🍳: No plan yet
🥚: Not started
🐣: Early Developing stage 🌱
🐥: Tests stage 🧪
🐤: v1 Completed 🎉
🐔: Advanced Developing stage🔨 (v2)