TCH AI NI Nirashanelki Why a Positional Encoding Layer in Transformers? Does the transformer models not know the positional data?
