jiasenlu / YOLOv3.pytorch

Pytorch implementation of Yolo V3

Pytorch implementation of Yolo V3.

This Code is still under development.

Introduction

This project is a pytorch implementation of Yolo v3, aimed to replicate the Darknet implementation. Recently, there are a number of implementations:

• qqwweee/keras-yolo3, developed based on Keras + Numpy
• eriklindernoren/PyTorch-YOLOv3, developed based on Pytorch + Numpy, loss does not coverge.
• ayooshkathuria/pytorch-yolo-v3, inference only, developed based on Pytorch + Numpy.
• BobLiu20/YOLOv3_PyTorch, Pytorch + Numpy, load pytorch pretrained model, loss does not converge now.

However, for Pytorch implementations, no one can replicate the performance of original darknet implementation. This project's goal is to benchmark the Yolo v3 in pytorch.

Benchmarking

Preparation

First of all, clone the code

git clone https://github.com/jiasenlu/YOLOv3.pytorch.git

Then, create a folder:

cd  YOLOv3.pytorch && mkdir data

prerequisites

• Pytorch 3.6
• Pytorch 0.5 (latest)
• TorchVision
• TensorBoard
• CUDA 8.0 or higher

Data Preparation

• PASCAL_VOC 07+12: Please follow the instructions in py-faster-rcnn to prepare VOC datasets. Actually, you can refer to any others. After downloading the data, creat softlinks in the folder data/.

• COCO: Please also follow the instructions in py-faster-rcnn to prepare the data.

• Visual Genome: Please follow the instructions in bottom-up-attention to prepare Visual Genome dataset. You need to download the images and object annotation files first, and then perform proprecessing to obtain the vocabulary and cleansed annotations based on the scripts provided in this repository.

Pretrained Model

We use converted pytorch pretrained model, you can place the pretrained model under data/weights.

• Darknet53: GoogleDrive
• Resnet101:
• Resnet50:
• MobelNet:

Compilation

Compile the cuda dependencies using following simple commands:

sh make.sh

Train

 python main.py

Test

Demo

Citation

@article{redmon2018yolov3,
  title={Yolov3: An incremental improvement},
  author={Redmon, Joseph and Farhadi, Ali},
  journal={arXiv preprint arXiv:1804.02767},
  year={2018}
}