Introduction

YOLOX is an anchor-free version of YOLO, with a simpler design but better performance! It aims to bridge the gap between research and industrial communities. For more details, please refer to our report on Arxiv.

This repo is an implementation of PyTorch version YOLOX, there is also a MegEngine implementation.

Updates!!

• 【2023/10/23】 We employ YOLOX(commit id ac58e0a5e68e57454b7b9ac822aced493b553c53) as the first stage in Apollo camera_detection_2d component.
• 【2023/02/28】 We support assignment visualization tool, see doc here.
• 【2022/04/14】 We support jit compile op.
• 【2021/08/19】 We optimize the training process with 2x faster training and ~1% higher performance! See notes for more details.
• 【2021/08/05】 We release MegEngine version YOLOX.
• 【2021/07/28】 We fix the fatal error of memory leak
• 【2021/07/26】 We now support MegEngine deployment.
• 【2021/07/20】 We have released our technical report on Arxiv.

Quick Start

Installation

Step1. Install YOLOX from source.

git clone git@github.com:LordonCN/YOLOX.git
cd YOLOX

conda create -n yolox python=3.8
conda activate yolox
pip3 install -r requirements.txt

Demo

Step1. Download a pretrained model from the benchmark table.

Model	size	Params (M)	Datasets	Class	weights
YOLOX-voc-s	640	26.8	KITTI	6	link
YOLOX-voc-s	640	26.8	L4-100w	8	link

Step2. Use either -n or -f to specify your detector's config. For example, here use best_L4_ckpt model:

python tools/demo.py image -n yolox-s -c /path/to/your/best_L4_ckpt.pth --path sample/ --conf 0.25 --nms 0.45 --tsize 640 --save_result --device [cpu/gpu]

then you will find result under path YOLOX_outputs/yolox_s/.

Reproduce our results on KITTI

Step1. Prepare KITTI dataset

cd <YOLOX_HOME>
ln -s /path/to/your/KITTI ./datasets/KITTI

Step2. Tools for kitti type datasets We provide tools for KITTI type datasets which can help to trans it to VOC type : readme

Step3. change kitti configs

• class number: 8 to 6

1. change voc_classes.py to KITTI class.
2. modify yolox_voc_s.py todo items.
3. modify voc.py line 119 change jpg to png

self._imgpath = os.path.join("%s", "JPEGImages", "%s.jpg") # to png

Step4. Reproduce our results on KITTI by specifying -n:

python3 tools/train.py -f exps/example/yolox_voc/yolox_voc_s.py -d 0 -b 16

or resume

python3 tools/train.py -f exps/example/yolox_voc/yolox_voc_s.py -d 0 -b 16 -c /path/to/your/latest_ckpt.pth --resume

• -d: number of gpu devices
• -b: total batch size, the recommended number for -b is num-gpu * 8
• --fp16: mixed precision training
• --cache: caching imgs into RAM to accelarate training, which need large system RAM.
• -c: checkpoint file path

Export

We support batch testing for fast evaluation:

python tools/export_onnx.py --input data -n yolox-s -c YOLOX_outputs/yolox_voc_s/latest_ckpt.pth  --output-name yolox.onnx

• --input: onnx model input blob name.
• -c: path of model.
• --output-name: the file name of covert model

Multi Machine Training

We also support multi-nodes training. Just add the following args:

• --num_machines: num of your total training nodes
• --machine_rank: specify the rank of each node

Suppose you want to train YOLOX on 2 machines, and your master machines's IP is 123.123.123.123, use port 12312 and TCP.

On master machine, run

python tools/train.py -n yolox-s -b 128 --dist-url tcp://123.123.123.123:12312 --num_machines 2 --machine_rank 0

On the second machine, run

python tools/train.py -n yolox-s -b 128 --dist-url tcp://123.123.123.123:12312 --num_machines 2 --machine_rank 1

Logging to Weights & Biases

To log metrics, predictions and model checkpoints to W&B use the command line argument --logger wandb and use the prefix "wandb-" to specify arguments for initializing the wandb run.

python tools/train.py -n yolox-s -d 8 -b 64 --fp16 -o [--cache] --logger wandb wandb-project <project name>
                         yolox-m
                         yolox-l
                         yolox-x

An example wandb dashboard is available here

Others

See more information with the following command:

python -m yolox.tools.train --help

Evaluation

We support batch testing for fast evaluation:

python -m yolox.tools.eval -n  yolox-s -c yolox_s.pth -b 64 -d 8 --conf 0.001 [--fp16] [--fuse]
                               yolox-m
                               yolox-l
                               yolox-x

• --fuse: fuse conv and bn
• -d: number of GPUs used for evaluation. DEFAULT: All GPUs available will be used.
• -b: total batch size across on all GPUs

To reproduce speed test, we use the following command:

python -m yolox.tools.eval -n  yolox-s -c yolox_s.pth -b 1 -d 1 --conf 0.001 --fp16 --fuse
                               yolox-m
                               yolox-l
                               yolox-x

Tutorials

• Training on custom data
• Caching for custom data
• Manipulating training image size
• Assignment visualization
• Freezing model

Deployment

1. MegEngine in C++ and Python
2. ONNX export and an ONNXRuntime
3. TensorRT in C++ and Python
4. ncnn in C++ and Java
5. OpenVINO in C++ and Python
6. Accelerate YOLOX inference with nebullvm in Python

Cite YOLOX

If you use YOLOX in your research, please cite our work by using the following BibTeX entry:

 @article{yolox2021,
  title={YOLOX: Exceeding YOLO Series in 2021},
  author={Ge, Zheng and Liu, Songtao and Wang, Feng and Li, Zeming and Sun, Jian},
  journal={arXiv preprint arXiv:2107.08430},
  year={2021}
}

In memory of Dr. Jian Sun

Without the guidance of Dr. Jian Sun, YOLOX would not have been released and open sourced to the community. The passing away of Dr. Jian is a huge loss to the Computer Vision field. We add this section here to express our remembrance and condolences to our captain Dr. Jian. It is hoped that every AI practitioner in the world will stick to the concept of "continuous innovation to expand cognitive boundaries, and extraordinary technology to achieve product value" and move forward all the way.

没有孙剑博士的指导，YOLOX也不会问世并开源给社区使用。 孙剑博士的离去是CV领域的一大损失，我们在此特别添加了这个部分来表达对我们的“船长”孙老师的纪念和哀思。 希望世界上的每个AI从业者秉持着“持续创新拓展认知边界，非凡科技成就产品价值”的观念，一路向前。