加速大模型训练的双向交错流水线并行方法
Bidirectional interleaved pipeline parallelism for accelerating large model training
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 8 流水线并行下的显存占用分布 (8 GPU) |
| Fig.8 Memory footprint distribution under pipeline parallelism (8 GPU) |
|
|