인터페이스는 현지화되어 있으며, Skill 소개와 설명 문서는 선택한 언어로 AI가 자동 번역합니다. 코드·명령어·경로·전문 용어는 원문을 유지하며 원작자 문서가 우선합니다.
Skills 둘러보기
개발 및 엔지니어링 · SKILL.md

moe-and-acceleration

moe-and-acceleration

Use this repo skill for Swin-Transformer optional Swin-MoE, Tutel, Apex, fused window-process CUDA extension, distributed checkpoint, and backend probe workflows.

Independent GitHub snapshot · Automated classification · Original documentation governs

Skill 문서

GitHub

name: moe-and-acceleration description: "Use this repo skill for Swin-Transformer optional Swin-MoE, Tutel, Apex, fused window-process CUDA extension, distributed checkpoint, and backend probe workflows." disable-model-invocation: true metadata: disco-role: operating license: MIT

moe-and-acceleration

Use this sub-skill when a task involves Swin-MoE, Tutel, optional Apex fused layernorm/optimizers, --fused_window_process, the CUDA window-process extension, or distributed MoE checkpoint behavior.

Covers

  • main_moe.py command shape and distributed expectations.
  • MODEL.SWIN_MOE config fields: experts, top-k, capacity factor, router type, aux loss, and MoE block placement.
  • Tutel dependency and why CPU is not a truthful substitute for MoE runtime verification.
  • Optional swin_window_process CUDA extension build/probe guidance.
  • Apex optional flags: fused layernorm, fused_adam, fused_lamb.
  • Rank-sharded MoE checkpoints and save/load caveats.

Routes elsewhere

  • Baseline Swin/SwinV2/Swin-MLP constructors: core-models.
  • Ordinary supervised main.py commands: training-eval-cli.
  • Generic checkpoint flag meaning and data layouts: data-and-checkpoints.
  • SimMIM commands: simmim-workflows.

Workflow

  1. Run scripts/check_optional_backends.py to see which optional backends are actually available.
  2. Read references/moe-and-acceleration.md for command and config rules.
  3. If a command is requested, build it with the root scripts/swin_cli_command_builder.py --workflow moe-train or moe-eval.
  4. If backend probes fail, read references/troubleshooting.md and avoid claiming runtime verification.

Verification boundary

CPU import checks are not evidence that Tutel MoE, Apex fused ops, or the custom CUDA extension works. Treat those as optional backend capabilities that require compatible CUDA/PyTorch/toolkit/runtime evidence.

Linked files

  • references/moe-and-acceleration.md - MoE and optional acceleration workflow details.
  • references/troubleshooting.md - missing Tutel/Apex/fused-kernel and distributed checkpoint issues.
  • scripts/check_optional_backends.py - safe optional backend probe.
저장한 Skills