teamdatatonic/tf-sampling

Effective sampling methods within TensorFlow input functions.

TensorFlow Sampling

Effective sampling methods within TensorFlow input functions.

Authors:

• William Fletcher
• Laxmi Prajapat

Table of Contents

• About the Project
  • Built With
  • Key Features
    • Sampling Techniques
    • Real-World Examples
• Getting Started
  • Installation
• Usage
  • Run Tests
• Contributing
• Licensing

About the Project

A collection of sampling techniques and real-world examples applied to training / testing data directly inside the input function using the tf.data API.

Built With

• Tensorflow (1.14)

Key Features

Sampling Techniques

There are two sampling methods presented:

• Downsampling with weight
• Negative sampling

Real-World Examples

Machine Learning examples on open-source datasets (local and AI Platform):

• Acquire Valued Shoppers
• Million Songs

Getting Started

Installation

• Clone this repository

git clone https://github.com/teamdatatonic/tf-sampling.git

• Installing the sampling module from source:

python setup.py install

Usage

Run Tests

Execute tests from sampling/tests/:

python -m pytest

Generate coverage report:

python -m pytest --cov=../io .

Contributing

If you'd like to contribute, please fork the repository and make changes as you'd like. Pull requests are welcome.

Licensing

Distributed under the MIT License. See LICENSE for more information.