Why the model I trained is latest.pdparams rather than best_accuracy.pdparams

[Haeppypuppy]

Please provide the following information to quickly locate the problem

when I used the pre_train model to train. However,the name of output model in C:\Users\PaddleOCR\output
is latest.pdparam rather than best_accuracy.pdparams

did I get any wrong step?

System Environment：Windows
Paddle： release 2.7
PaddlePaddle: 2.3.1

Command Code：python tools/train.py -c configs/det/det_mv3_db.yml -o Global.pretrained_model=./pretrain_models/MobileNetV3_small_x0_35_pretrained
Complete Error Message：NA

We provide AceIssueSolver to solve issues, do you want it? (Please write yes/no): Yes

Please try to not include the image in the issue.

[cuicheng01]

1. Please make sure you have actually used the eval data set
2. Please make sure you have actually performed the eval operation

[Haeppypuppy]

I found the setting and problem.

In program.py

save_model( model, optimizer, save_model_dir, logger, config, is_best=True, # prefix='best_accuracy', # 將prefix設置為'best_accuracy' best_model_dict=best_model_dict, epoch=epoch, global_step=global_step)
is_best=True, # True ==save the best model when training
is_best=False, # False ==save the latest model when training