Official NVIDIA-authored guidance for PhysicsNeMo ShardTensor domain parallelism — integrate domain parallelism into training/inference scripts (new or existing) with DDP or FSDP2, write and register shard patches to enable new layers/ops, and bootstrap multi-GPU correctness tests.
安装
npx skills add NVIDIA/skills --skill physicsnemo-shard-tensornpx skills add NVIDIA/skills --skill physicsnemo-shard-tensor -a claude-code -gnpx skills add NVIDIA/skills --skill physicsnemo-shard-tensor -a codex -ggemini skills install https://github.com/NVIDIA/skills.git --path skills/physicsnemo-shard-tensorgh skill install NVIDIA/skills physicsnemo-shard-tensorgit clone --depth 1 https://github.com/NVIDIA/skills /tmp/NVIDIA__skills
mkdir -p ~/.claude/skills ~/.agents/skills
cp -R /tmp/NVIDIA__skills/skills/physicsnemo-shard-tensor ~/.claude/skills/physicsnemo-shard-tensor # Claude Code
cp -R /tmp/NVIDIA__skills/skills/physicsnemo-shard-tensor ~/.agents/skills/physicsnemo-shard-tensor # Codex, Cursor, Gemini CLI, Copilot, OpenCode