a machine learning project management

Thanks for the original contribution from the artical Version your machine learning models with Sacred and Managing Machine Learning projects.

This example reorganize the project monchewharry/ML-manage-tfexample into general project structure as suggested by the mentioned credit above. The project running is similar, just by the following and read the printted hints to start omniboard.

python run.py

todo

• add hyperparameter tune (no duplicate experiment control)
• add unittest
• add docker

structure

.
├── README.md
├── run.py
├── src
│  ├── config.py
│  ├── data_processing
│  │  └── data_loader.py
│  ├── inference.py
│  ├── models
│  │  └── model.py
│  ├── train.py
│  ├── utils
│  └── visualization
│     └── explore.py
└── tests
   ├── data
   │  └── test-set.py
   ├── integration
   │  └── test_model.py
   └── unit
      └── test_data_loader.py

More

• mongodb macOS
• sacred example
• omniboard quick-start