Welcome to

Luminoth is an open source toolkit for computer vision. Currently, we support object detection, but we are aiming for much more. It’s built in Python, using TensorFlow.

The code is open source and available on GitHub.

Documentation

Usage:

• Installation
  • Before you start
  • Installing from PyPI
  • Installing from source
• Getting started
• Tutorial
  • First steps
  • Building custom traffic dataset
  • Training the model
  • Using TensorBoard to visualize the training process
  • Evaluating models
  • Creating and sharing your own checkpoints
  • Using Luminoth from Python
• Adapting a dataset
  • Conversion tool
  • Merge tool
• Training your own model
  • Configuration
  • Training
• Evaluating a model
• Cloud management
  • Pre-requisites
  • Running a training job
  • Resuming a previous training job
  • Listing jobs
  • Fetching logs
  • Running an evaluation job
  • Results
• Working with checkpoints

API reference:

• Datasets
• Models

CLI reference:

• Checkpoint management
• Cloud management
• Dataset management
• Evaluating a model
• Predict with a model
• Web server
• Training a model