Tensorflow-Scikit Workflows

Demonstration of a project structure for common machine learning workflows. Uses Tensorflow and Scikit with a monorepo structure.

An example of executing via REST (FastAPI) is also included.

Setup and Install

Create a Virtual Environment

$ pyenv virtualenv 3.11.9 tf-sk-workflows

Activate Existing Virtual Environment

$ pyenv local tf-sk-workflows

Install

poetry install

Project Structure

core Core classes and abstractions related to all models, representations of model performance and comparison utils.

sklearnworkflows Machine learning pipelines implemented using Sklearn.

tensorflowworkflows Machine learning pipelines implemented using Tensorflow.

api A simple FastAPI interface that exposes some API endpoints to demonstrate executing the workflows via REST.

API and Swagger Spec

Run API Server

poetry run python api

Swagger Spec

Available at:

SwaggerUI: http://localhost:3001/docs

Redoc: http://localhost:3001/redoc

JSON Spec: http://localhost:3001/openapi.json