Generating Synthetic Audio Data for Attention-Based Speech Recognition Systems