PySubmarine

PySubmarine is aiming to ease the ML engineer's life by providing a set of libraries.

It includes a high-level out-of-box ML library like deepFM, FM, etc. low-level library to interact with submarine like creating experiment, tracking experiment metrics, parameters.

Package setup

• Install latest version of pysubmarine

git clone https://github.com/apache/submarine.git
cd submarine/submarine-sdk/pysubmarine
pip install .

# install extras_require
## bash
pip install -e .[tf2,pytorch]
## zsh
pip install -e ".[tf2,pytorch]"

• Install package from pypi

pip install apache-submarine

Easy-to-use model trainers

• FM
• DeepFM

Submarine experiment management

Makes it easy to run distributed or non-distributed TensorFlow, PyTorch experiments on Kubernetes.

• mnist example

Development

See Python Development in the documentation subproject.