zlszhonglongshen / centernet-1

A resnet18 version of CenterNet（objects as points）

Update

Recently, I publish a better CenterNet, called CenterNet-plus. Although my CenterNet-plus has a very simple piepline with any DCN layers, it surpasses the official CenterNet.

You can get my CenterNet-plus from the following project:

https://github.com/yjh0410/CenterNet-plus

CenterNet

A PyTorch version of CenterNet（objects as points）. I only support resnet18 version. No DLA or Hourglass version.

I have trained it on VOC0712 and COCO 2017. You can download them from BaiDuYunDisk：

Link：https://pan.baidu.com/s/170OYftGRVW-j5qAKYyHSQQ

Password：jz4q

The official CenterNet takes advantage of DCN while I just replace it with SPP used in YOLOv3 as I'm a little lazy ~

On VOC:

	data	mAP
(official) resnet18 + DCN	VOC2007	75.7
(Our) resnet18 + SPP	VOC2007	75.6

On COCO:

	data	AP	AP50
(official) resnet18 + DCN	COCO val	28	44.9
(Our) resnet18 + SPP	COCO val	25.8	45.4

I'm still trying something new to make my CenterNet-Lite stronger.

Installation

• Pytorch-gpu 1.1.0/1.2.0/1.3.0
• Tensorboard 1.14.
• opencv-python, python3.6/3.7

Dataset

VOC Dataset

I copy the download files from the following excellent project: https://github.com/amdegroot/ssd.pytorch

I have uploaded the VOC2007 and VOC2012 to BaiDuYunDisk, so for researchers in China, you can download them from BaiDuYunDisk:

Link：https://pan.baidu.com/s/1tYPGCYGyC0wjpC97H-zzMQ

Password：4la9

You will get a VOCdevkit.zip, then what you need to do is just to unzip it and put it into data/. After that, the whole path to VOC dataset is data/VOCdevkit/VOC2007 and data/VOCdevkit/VOC2012.

Download VOC2007 trainval & test

# specify a directory for dataset to be downloaded into, else default is ~/data/
sh data/scripts/VOC2007.sh # <directory>

Download VOC2012 trainval

# specify a directory for dataset to be downloaded into, else default is ~/data/
sh data/scripts/VOC2012.sh # <directory>

MSCOCO Dataset

I copy the download files from the following excellent project: https://github.com/DeNA/PyTorch_YOLOv3

Download MSCOCO 2017 dataset

Just run sh data/scripts/COCO2017.sh. You will get COCO train2017, val2017, test2017.

Train

VOC

python train.py --cuda -d voc

You can run python train.py -h to check all optional argument.

COCO

python train.py --cuda -d coco

Test

VOC

python test.py --cuda -d voc --trained_model [ Please input the path to model dir. ]

COCO

python test.py --cuda -d coco-val --trained_model [ Please input the path to model dir. ]

Evaluation

VOC

python eval.py --cuda -d voc --train_model [ Please input the path to model dir. ]

COCO

To run on COCO_val:

python eval.py --cuda -d coco-val --train_model [ Please input the path to model dir. ]

To run on COCO_test-dev(You must be sure that you have downloaded test2017):

python eval.py --cuda -d coco-test --train_model [ Please input the path to model dir. ]

You will get a .json file which can be evaluated on COCO test server.