1
BigMac: Breaking the Pareto Frontier of Compute and Memory in Multimodal LLM Training
突破多模态LLM训练中计算与内存的帕累托前沿，BigMac提出高效优化方案。
arXiv:2605.25451v1 Announce Type: new Abstract: Training multimodal large language models (MLLMs) is challenged by both model and data heterogeneity. …