sk Skill · onescience-ai
onescience-parallel
将 PyTorch 模型改造为支持 Pipeline Parallelism (PP) + Tensor Parallelism (TP) 的分布式训练模型。 适用场景:单机单卡模型改多卡分布式、模型拆分为多个 pipeline stage、替换 nn.Linear 为并行线性层、 编写 forward_step_func / model_provider / dataset_provider、配置 PipelineTensorShapeConfig、 创建 Distributed 版本模块(DistributedFuser / DistributedAttention / DistributedMlp)。 当用户提到以下任何关键词时,务必使用此 skill: "流水线并行"、"pipeline parallel"、"模型并行改造"、"分布式训练"、"stage 拆分"、…
Open on skills.sh ↗read 2026-09-17
- installs 8w
- 0
- 30-day movement
- starts with the next reading
- Related entries
- 1
- Connections
- 0
pythonbashPython
- Host repository
- onescience-ai/OneSkills
- Host stars
- 20
- Host language
- Python