This is just a simple implementation of "CenterNet:Objects as Points".
office: https://github.com/xingyizhou/CenterNet

Unlike the other repo, I just used pure pytorch. Unlike most implementation base on detectron2 or mmdetection, I made the code as simple as possible to understand, no trick, as close to the original performance as possible(I hope so).
Anyway, if someone looks at the code and sees any bugs, feel free to let me know.

• Be able to train smoothly and converge.
• Inference.
• COCO and PASCAL_VOC evaluate.
• Improve performance.
• More backbone.
• So on.

Win10 or Linux.

torch >= 1.2.0  
torchvision>=0.4.1  
timm >= 0.1.14
Pillow >= 6.2.2
opencv-python >= 4.2.0
albumentations >= 0.4.5

Modify config/{}.py according to your needs, then python train_{}.py.

Modify this script inference.py that is easy to understand, then python inference.py.