PaddleX
PaddlePaddle End-to-End Development Toolkit
..., the validation set and the test set. Therefore, we need to divide the above data. Using the paddlex command, the data set can be randomly divided into 70% training set, 20% validation set and 10% test set. If you use the PaddleX visualization client for model training, the data set division function is integrated in the client, and you do not need to use command division by yourself.