Results for “expert-parallelism”
62 skillsmoe-training
Train Mixture of Experts (MoE) models using DeepSpeed or HuggingFace, covering architectures, routing, load balancing, and expert parallelism.
10.4k · bundle
training-llms-megatron
Trains large language models (2B-462B parameters) using NVIDIA Megatron-Core with advanced parallelism strategies for maximum GPU efficiency.
10.4k · bundle
nemo-mbridge-perf-moe-long-context
Provides guidance for training Mixture-of-Experts models with long context windows, covering context parallelism sizing, selective recomputation, dispatcher choices, and practical patterns from recent experiments.
2.2k · bundle
deepspeed
Provides expert guidance for distributed training with DeepSpeed, covering ZeRO optimization stages, pipeline parallelism, FP16/BF16/FP8, 1-bit Adam, and sparse attention.
10.4k · bundle
deepspeed
Expert guidance for distributed training with DeepSpeed - ZeRO optimization stages, pipeline parallelism, FP16/BF16/FP8, 1-bit Adam, sparse attention
1 · bundle
More results
deepspeed
Expert guidance for distributed training with DeepSpeed - ZeRO optimization stages, pipeline parallelism, FP16/BF16/FP8, 1-bit Adam, sparse attention
0 · bundle
training-llms-megatron
Trains large language models (2B-462B parameters) using NVIDIA Megatron-Core with advanced parallelism strategies. Use when training models >1B parameters, need maximum GPU efficiency (47% MFU on H100), or require tensor/pipeline/sequence/context/expert parallelism. Production-ready framework used for Nemotron, LLaMA, DeepSeek.
1 · bundle
parallel-phases
Execute phased plans with multiple independent tasks per phase by fanning out one agent per task, reconciling outcomes per wave, gating between phases with verify commands, and emitting a phase × outcome report. Triggers include "execute this plan", "work through these phases", "swarm over this backlog", "parallelize this plan".
1 · bundle
anonymous-expert
Embody the Anonymous hacktivist collective as an AI persona, using integrated methodology skills for leaderless coordination, memetic warfare, swarm operations, and collective identity.
6
pr-review-expert
Resolves legacy references to the pr-review-expert capability by routing to the current runtime agent or skill.
20
external-context
Invoke parallel document-specialist agents for external web searches and documentation lookup
1
plan-arbiter
Compare, cross-review, and merge competing plans from multiple agents into one executable direction with a clear handoff.
3.4k · bundle
build-parallelism
Optimize MSBuild build parallelism by configuring /maxcpucount, graph build mode, project references, and analyzing binlogs to reduce multi-project solution build times.
4k
elon-musk-expert
Provides a persona that applies Elon Musk's engineering and management principles, including first principles reasoning, the 5-step algorithm, and physics-based analysis, to problem-solving and decision-making.
6
parallel-agents
Multi-agent orchestration patterns. Use when multiple independent tasks can run with different domain expertise or when comprehensive analysis requires multiple perspectives.
2
cogvlm-visual-expert-for-pretrained-language-models-arxiv-23
CogVLM: Visual Expert for Pretrained Language Models
6
parallel-agents
Multi-agent orchestration patterns. Use when multiple independent tasks can run with different domain expertise or when comprehensive analysis requires multiple perspectives.
0
bmad-ml-astra
Interdisciplinary synthesis and cross-domain transfer specialist. Use when the user asks to talk to Astra, requests the synthesizer, or needs cross-field insight connections and multi-modal research directions.
0 · bundle
bob-hope-expert
Embody the comedic persona of Bob Hope, delivering topical, self-deprecating, and bipartisan humor with impeccable timing and adaptability to any audience.
6
vercel-react-expert
Resolves legacy references to the vercel-react-expert capability by routing to the current runtime agent, plugin, or narrower skill.
20
al-ghazali-expert
Embody the persona of Al-Ghazali to provide spiritual guidance, philosophical critique, and practical remedies for spiritual ailments.
6
ultrawork
Parallel execution engine for high-throughput task completion
1
expert-security
安全专家入口。用于 Codex CLI 的 $expert-security 调用。 适用于威胁建模、漏洞评估、安全代码审查、安全架构设计、DevSecOps、安全运营、事件响应、合规审计和完整安全健康评估。 触发词:安全专家、威胁建模、STRIDE、OWASP、SAST、DAST、SBOM、漏洞评估、代码审计、事件响应、合规审计、SOC、等保、GDPR、PIPL、隐私政策审查
0 · bundle
test-parallelizer
Test Parallelizer - Auto-activating skill for Test Automation. Triggers on: test parallelizer, test parallelizer Part of the Test Automation skill category. Use when writing or running tests. Trigger with phrases like "test parallelizer", "test parallelizer", "test".
4
subagent-result-merge
Merges agent or review outputs into one deduplicated, evidence-linked, severity-ranked actionable report.
0 · bundle
dispatching-parallel-agents
Dispatch multiple independent tasks to parallel agents for faster debugging and problem-solving.
247k
bo-burnham-expert
Embody the voice and methodology of comedian Bo Burnham, blending meta-commentary, musical comedy, and cultural critique in responses.
6
larry-page-expert
Adopts the voice and methodology of Larry Page to reframe problems with 10x thinking, moonshot criteria, and long-term bets.
6
expert-team
专家团总路由器。用于 Codex CLI 的 $expert-team 调用。 自动在软件开发团队、设计原型专家团、产品战略团队、基础设施运维专家、安全专家和数据库优化专家之间路由,也支持显式指定 software/design/product/ops/security/database。 触发词:专家团、团队协作、软件开发、设计原型、产品战略、基础设施运维、安全专家、数据库专家、威胁建模、代码审计、SRE、PRD、竞品、路线图、监控、部署、安全加固、SQL、索引、慢查询、迁移
0 · bundle
karl-marx-expert
Embody Karl Marx's voice and methodology for dialectical, historically grounded analysis of capitalism, class, and social change.
6
guy-debord-expert
Adopts the voice and methodology of Guy Debord to analyze content through spectacle, détournement, and refusal, offering subversive critiques and creative transformations.
6
dispatching-parallel-agents
Dispatching Parallel Agents
0
dalai-lama-expert
Embody the Dalai Lama's voice and teachings to provide compassionate guidance, secular ethics, and meditation practices.
6
agent-hub
Multi-agent collaboration plugin that spawns N parallel subagents competing on the same task via git worktree isolation. Agents work independently, results are evaluated by metric or LLM judge, and the best branch is merged. Use when: user wants multiple approaches tried in parallel — code optimization, content variation, research exploration, or any task that benefits from parallel competition. Requires: a git repo.
65 · bundle
ultrawork
Parallel execution engine for high-throughput task completion
1
research-analyst
Conducts thorough landscape research, competitive analysis, best practices evaluation, and evidence-based recommendations. Expert in market research and trend analysis.
10