Expert guidance for distributed training with DeepSpeed - ZeRO optimization stages, pipeline parallelism, FP16/BF16/FP8, 1-bit Adam, sparse attention
安装
npx skills add davila7/claude-code-templates --skill deepspeednpx skills add davila7/claude-code-templates --skill deepspeed -a claude-code -gnpx skills add davila7/claude-code-templates --skill deepspeed -a codex -ggemini skills install https://github.com/davila7/claude-code-templates.git --path cli-tool/components/skills/ai-research/distributed-training-deepspeedgh skill install davila7/claude-code-templates deepspeedgit clone --depth 1 https://github.com/davila7/claude-code-templates /tmp/davila7__claude-code-templates
mkdir -p ~/.claude/skills ~/.agents/skills
cp -R /tmp/davila7__claude-code-templates/cli-tool/components/skills/ai-research/distributed-training-deepspeed ~/.claude/skills/deepspeed # Claude Code
cp -R /tmp/davila7__claude-code-templates/cli-tool/components/skills/ai-research/distributed-training-deepspeed ~/.agents/skills/deepspeed # Codex, Cursor, Gemini CLI, Copilot, OpenCode