Number of training steps for fine-tuning BERT's language model

Question

I want to fine-tune BERT-base-uncased for the language model, according to my custom dataset. I was over the internet in last couple of days, but found nothing useful for my