Understanding transformer models for NLP tasks

I've been studying transformer architecture and while I understand the basics of attention mechanisms, I'm struggling with some concepts. Can someone explain: 1. Why positional encoding is necessary? 2. The difference between self-attention and cross-attention? 3. How to choose between encoder-only, decoder-only, and encoder-decoder models? Any good resources for deep-diving into transformers would be appreciated!