How do I get Precision and Recall to be calculated together when I run model_main_tf2.py ?

Hello, I usually create my models by using Tensorflow2 Object Detection API and my original data.

When I create a model, I always only see the Loss graph in Tensorboard.
I know that if I run model_main_tf2.py with the option "--checkpoint_dir" after creating the model, Precison and Recall will be calculated.
However, I want to check the progress of learning while calculating Precision and Recall during running model_main_tf2.py.

I remember when I used Tensorflow-gpu1.15 before, the training was going on while Precision and Recall were being calculated.
If Precision and Recall are not computed at the same time as learning, I think it means that Cross Validation is not performed when creating a model in Tensorflow2.

If you know how to train and create a model while calculating Precision and Recall when running model_main_tf2.py, please let me know.

Thanks!

・OS : Ubuntu20.04LTS
・NVIDIA Driver : 450.102
・CUDA : 11.0
・cuDNN : 8.0.2
・Python : 3.6
・Tensorflow-gpu : 2.4.1