johnolafenwa / deepstack-trainer

Custom Object Detection Training for DeepStack

DeepStack Custom Object Detection Training

DeepStack is a cross platform AI engine for performing Object Detection, Face Detection and Face Recognition on the edge and the cloud.

This repo provides functionality to train object detection models on your own objects, the model from this can be instantly deployed to DeepStack on Desktop devices, Nvidia Jetson devices and the cloud.

Follow the instructions in this repo for setup, preparing your dataset, training your model and deploying for production use with DeepStack.

Setup and Train on Google Colab

Local Setup and Installation

Clone DeepStack Trainer

git clone https://github.com/johnolafenwa/deepstack-trainer

Install Requirements

• Install Python3.7 or higher
• Install PyTorch
• Run

   pip install -r requirements.txt