中文文档:README
A Semi-automatic and Automatic Annotation Toolkit for cvat
First of all you should download the project CVAT and deploy to local
Install according to the guide CVAT && nuclio
After the above has been completed
- Put the training yolov4 weight file (train with darknet)「 *.weights 」and config file「 *.cfg 」into directory yolo-weight
- Build this project into docker image and then you can use for automatic annotation
Because I created the project under the project CVAT so the docker image is
./deploy_cpu.sh ./openvino/omz/public/yolo-v4-tf/
If you find that the NODE PORT is 0, then you have to check your build if there is something wrong.
Normally, there will be two output as below
Below is an example, single and multiple product image recognition.