What is validation_step argument?

I searched for the description of –validation_step argument in util/flags.py code but it is not there.
I also run ./DeepSpeech --helpfull, but it is not listed there.

What is it for?

My training is not saving a dev step_loss summary results and don’t know why! Any suggestions would be much appreciated.

This was used to set the recurrence of running WER computation against the validation set.

Can you be more specific ?

Thank you for the prompt reply.

I mean I am passing summary_dir parameter to my run, then I am visualizing the summary results using tensorboard:
tensorboard --logdir=$summary_dir

Their is a plot for train, with a scalar step_loss, but not for dev/develpment

I have just tried with --validation_step 1, then the dev plot appeared!