DEtection TRansformer Network (DETR) - TF2

Overview

Implementation of the DETR (DEtection TRansformer) network (Carion, Nicolas, et al., 2020) in Tensorflow 2.

This is still a work in progress!

References

• Research Paper: End-to-End Object Detection with Transformers
• Original PyTorch Implementation: GitHub

Getting started

It is recommended to create a virtual environmet using Anaconda and Pip.

1. Create and start conda environmet:

conda create -n detr python=3.6 anaconda pip -y
conda activate detr

2. Install necessary libraries:

conda install tensorflow-gpu==2.x -y
pip install tensorflow-gpu==2.3.0rc1 tensorflow-addons==0.11.1 numpy scipy matplotlib imgaug opencv-python

3. Use the train_detr-tf2.ipynb notebook to start training and make inferences. The folder kangaroo contains a small dataset so you can start to work with it.

Code organization

In progress...

Known issues

• The model is currently not converging to good results. It may be due to the implemented loss function. So far, many investigations are beeing done around it.