Language Translation
摘要
This chapter explores modern machine translation techniques using LLMs such as MarianMT, mBART, T5, and mT5, and SLMs like DistilBART, TinyBERT, and MiniLM. It discusses the complexities of translating across languages—covering syntax, semantics, idioms, and domain-specific jargon—while comparing the strengths of LLMs and SLMs in terms of quality, latency, and computational cost.