Using Deep Speech

As suggested @SamahZaro , I moved the question to Language Model during training effect and @lissyx saw it and was kind enough to give some inputs. Thanks again.