Which is the way to monitoring the training process with Tensorboard?

[trungtv]

Is there any guideline on how to log training metrics to Tensorboard? Is there something similar to callback mechanism in SB 3?

[Yonv1943]

It is possoble to add metrics to TensorBoard:

with test_summary_writer.as_default():
    tf.summary.scalar('loss', test_loss.result(), step=epoch)

ElegantRL use the following methods to print and plot the training logging metrics https://github.com/AI4Finance-Foundation/ElegantRL/issues/128#issuecomment-1079674617

You can log training metrics by change the following code: (See the log_tuple, which is the return of agent.update_net(...)) https://github.com/AI4Finance-Foundation/ElegantRL/blob/a082978782aa9c2ef7956cda445db5817e4734e9/elegantrl_helloworld/run.py#L293-L302