加速大模型训练的双向交错流水线并行方法
吴厚明,陈岭

Bidirectional interleaved pipeline parallelism for accelerating large model training
Houming WU,Ling CHEN
图 10 流水线-数据并行下的显存占用分布(32 GPU)
Fig.10 Memory footprint distribution under pipeline-data parallelism (32 GPU)