Megalodon: Efficient LLM Pretraining and Inference with Unlimited Context LengthXuezhe Ma,Xiaomeng Yang,Wenhan Xiong,Beidi Chen,LILI YU,Hao Zhang,Jonathan May,Luke Zettlemoyer,Omer Levy,Chunting ZhouNeurIPS 2024(2024)引用 31|浏览145关键词Mega,Efficient Architecture,Long Sequence Modeling,Unlimited Context LengthAI 理解论文溯源树样例生成溯源树,研究论文发展脉络Chat Paper正在生成论文摘要