Meet LLaSM: An End-to-End Trained Large Multi-Modal Speech-Language Model with Cross-Modal Conversational Abilities Capable of Following Speech-and-Language Instructions