Skip to content

Explore skills

Search every indexed skill by name, description, repository or instructions, then narrow the results.

Showing 56971–57000 of 63406 skills

ascend-cluster-fast-slow-rank-detectorAscend/msagent专门用于 Ascend 集群 Profiling 性能数据的“快慢卡”诊断专家技能。当用户提供【集群性能数据目录/路径】并要求分析【快慢卡】、【慢节点】、【负载不均衡】或【集群瓶颈】时,必须触发此技能。该技能会自动接收集群路径,调度相关工具输出快慢卡的宏观定性与微观根因(如 Host 下发瓶颈、算子计算劣化)。Stars: 0Updated Claude CodeCodexUnverifiedInstall
ascend-communication-analysisAscend/msagentAnalyze Ascend NPU collective communication profiling data with a DB-first workflow. Use when the user provides `cluster_analysis_output/cluster_analysis.db`, rank-level `analysis.db`, rank-level `ascend_pytorch_profiler_{rank_id}.db`, together with `profiler_info.json`, and asks about HCCL or hcom communication cost, collective communication TOP ops, wait time, slow rank/straggler, Notify Wait, bandwidth, retry, relay, SDMA/RDMA/HCCS links, communication matrix, or Ascend communication fault patterns.Stars: 0Updated Claude CodeCodexUnverifiedInstall
ascend-computation-analysisAscend/msagentAnalyze Ascend NPU computation-side profiling data for single-card runs or a selected rank from multi-card runs. Use this skill when the user asks to diagnose computation bottlenecks, AI Core / AI Vector / AICPU hotspots, dynamic shape overhead, block dim issues, redundant TransData/Transpose/Cast, cross-stream waits, frequency/runtime-state issues, or fusion opportunities.Stars: 0Updated Claude CodeCodexUnverifiedInstall
ascend-fusion-op-replaceAscend/msagent针对昇腾 NPU 上的 PyTorch 模型,识别可做融合算子替换的代码段、替换为 torch_npu 融合算子并做性能/精度验证。当用户提到融合算子、NPU 算子替换、RMSNorm/SwiGLU/RoPE/Attention 融合、torch_npu 亲和算子等关键词是,使用本 Skill。Stars: 0Updated Claude CodeCodexUnverifiedInstall
ascend-msprof-analyze-cliAscend/msagentMindStudio Profiler Analyze(msprof-analyze)是面向 AI 训练与推理场景的性能分析工具,基于采集得到的 profiling 数据进行统计、比对和诊断,帮助定位计算、通信、调度及集群场景下的性能瓶颈。Stars: 0Updated Claude CodeCodexUnverifiedInstall
ascend-npu-snapshot-analyzerAscend/msagent分析 PyTorch memory snapshot pickle 文件(_dump_snapshot 导出的内存快照),提供内存峰值、碎片、泄漏、OOM 检测与交互式 HTML 报告。仅当用户明确提及 memory snapshot、内存快照、_dump_snapshot、pickle 内存文件 或需要对 snapshot pickle 做内存分析时才触发,不处理一般的 NPU 性能分析、SQLite 查询或通用内存问题。Stars: 0Updated Claude CodeCodexUnverifiedInstall
ascend-prof-compareAscend/msagentCompares two Prof datasets (GPU vs NPU / NPU vs NPU) by auto-installing and calling msprof-analyze compare, then parses the result xlsx into HTML report and Chinese xlsx. Invoke when user provides two prof dirs to compare, or has a compare result xlsx to analyze/translate.Stars: 0Updated Claude CodeCodexUnverifiedInstall
ascend-profiler-collect-adaptionAscend/msagentAdapt torch_npu.profiler interface for AI frameworks and training/inference code on Ascend NPU. Use this skill whenever users need to integrate profiling, add performance data collection, or instrument PyTorch code with torch_npu.profiler on Huawei Ascend NPU platforms. Triggers on: "torch_npu profiler", "NPU profiling", "Ascend profiler", "profiler adaptation", "profile training", "profile inference", "add profiler to framework", "性能采集", "profiling 数据", "昇腾 profiler", and any request to add performance monitoring to Ascend NPU code.Stars: 0Updated Claude CodeCodexUnverifiedInstall
ascend-profiler-data-validationAscend/msagent当用户提供 MindStudio profiler 采集的性能数据(框架 profiler、msprof 命令行)时,对数据完整性、采集状态及关键配置进行校验,确保后续分析工具能正常运行。Stars: 0Updated Claude CodeCodexUnverifiedInstall
ascend-profiler-db-explorerAscend/msagent面向 Ascend PyTorch Profiler / msprof DB(如 ascend_pytorch_profiler*.db、msprof_*.db)的 SQL 分析技能。将自然语言问题(算子耗时、通信、下发、调度、schema/table 查询)转为安全可执行 SQL,并按需从官方文档提取表结构详情。Stars: 0Updated Claude CodeCodexUnverifiedInstall
ascend-schedule-analysisAscend/msagentAnalyze Ascend NPU schedule, operator dispatch, operator launch, and Host Bound profiling issues in Ascend profiling data. Use when need to diagnose device Free time, framework/operator dispatch latency, launch latency, PYTORCH_API/CANN_API launch gaps, aclrtSynchronizeStream stalls, task queue behavior, CPU scheduling interference, GC/lock pauses, CPU affinity, or schedule-side optimization actions.Stars: 0Updated Claude CodeCodexUnverifiedInstall
memory-analysisAscend/msagent对 msMemScope 显存数据做全链路处理:根据用户的显存问题诉求(OOM、显存泄漏、显存占比拆解、Step 间对比等),指导并帮助用户以最合适的 msMemScope 采集配置完成显存数据采集(支持命令行、Python 接口、mstx 打点三种采集方式,覆盖训练、推理 vLLM-Ascend、FSDP、强化学习 verl 等场景);对已采集的 memscope_dump_*.csv/db 数据做系统性解读与问题诊断(整体显存曲线趋势分析、显存使用大头按模块/组件拆解分析、显存峰值点各模块与流程占比分析,以及对显存泄漏、OOM 等典型问题的根因定位与原因分析)。Stars: 0Updated Claude CodeCodexUnverifiedInstall
mindstudio-cpu-bindingAscend/msagentUse when diagnosing NPU + PyTorch or LLM Serving Host CPU affinity, NUMA locality, cgroup/cpuset constraints, CPU range conflicts, PyTorch/runtime threading, DataLoader, tokenizer, scheduler, vLLM-Ascend, SGLang, TTFT, TPOT, tokens/s, QPS, or multi-rank/multi-worker CPU binding issues.Stars: 0Updated Claude CodeCodexUnverifiedInstall
msagent-profiler-breakdownAscend/msagentDB格式Profiling数据场景化性能拆解:自动完成场景确认(框架/任务类型判定)、阶段拆解与单次执行边界拆解,并给出报告。Use when 需要对昇腾 Profiling DB 做阶段拆解、step 归因或生成性能报告, or when the user says "帮我按任务阶段拆解这个Profiling数据", "多层次性能耗时拆解", "分析这个 vllm/sglang/verl 数据,主要耗时都花费在哪里了"。Stars: 0Updated Claude CodeCodexUnverifiedInstall
op-mfu-calculatorAscend/msagent计算算子(如 matmul/GEMM)的 MFU(Machine FLOP Utilization),并给出清晰的公式和推导过程。Stars: 0Updated Claude CodeCodexUnverifiedInstall
op-mfu-profilerAscend/msagent运行程序采集 Profiling 数据,再用 msprof-analyze 分析TFLOPS 与 MFU。Stars: 0Updated Claude CodeCodexUnverifiedInstall
gen-evaluation-cfgAscend/msagentGenerate msmodelslim evaluation YAML configuration (service_oriented + aisbench + vllm-ascend). Use when user asks for evaluation config generation.Stars: 0Updated Claude CodeCodexUnverifiedInstall
msmodelslim-adapter-verificationAscend/msagent为 msModelSlim 适配器执行功能性验证。适用于基础适配器开发完成后,自动执行四步验证(测试模型、全回退量化、权重一致性与可加载/保存、实际量化规则校验)并输出通过/失败结论。Stars: 0Updated Claude CodeCodexUnverifiedInstall
msmodelslim-anti-outlier-adaptAscend/msagent在基础模型适配完成后,按算法适配并安装 Adapter 接口,依据目标模型真实 forward 生成并验证 final-logits patch,执行单算法 msModelSlim 离群值抑制变换、官方计算图提取和浮点 logits 门禁。Stars: 0Updated Claude CodeCodexUnverifiedInstall
msmodelslim-ep-parallel-adaptationAscend/msagentStars: 0Updated Claude CodeCodexUnverifiedInstall
msmodelslim-layer-wise-quantizationAscend/msagent为 msModelSlim 适配器实现逐层量化(按层加载/懒加载)能力。仅在用户明确要求逐层量化或基础适配因 CPU 内存不足无法全量加载权重时使用。该特性为高阶可选项,不是基础适配必需项。Stars: 0Updated Claude CodeCodexUnverifiedInstall
msmodelslim-model-adaptAscend/msagent为 msModelSlim 创建基础 Transformers 模型适配器(Model Adapter)。 包含创建适配器、实现必需接口与注册安装流程。 适用:Decoder-only LLM、理解类 VLM(仅 LLM/text 部分)。 多模态生成(DiT/扩散,含双/多专家)见下方「多模态生成扩展」节,按需加载。 不适用:Encoder-only、非 Transformers 架构。Stars: 0Updated Claude CodeCodexUnverifiedInstall
msmodelslim-model-analysisAscend/msagentStars: 0Updated Claude CodeCodexUnverifiedInstall
msmodelslim-model-dequantAscend/msagent为 msModelSlim 适配流程注入反量化能力。先识别模型权重是否可反量化,再实现反量化脚本并接入 model_adapter。当前仅覆盖 FP8 的 per-block 与 per-channel 两类;若格式不确定或无公开反量化规则,要求用户提供反量化脚本或浮点权重。Stars: 0Updated Claude CodeCodexUnverifiedInstall
msmodelslim-quick-quantAscend/msagent提供 msModelSlim 的通用快速量化指引,包含安装、最简 YAML 配置与基础执行校验。适用于用户询问 msmodelslim 安装、快速量化、配置 yaml、linear_quant 或 minmax 基础参数时。Stars: 0Updated Claude CodeCodexUnverifiedInstall
quant-tuning-evaluateAscend/msagent执行模型测评(统一入口)。LLM/VLM 走主流程 scripts/run_evaluation.py + Evaluation YAML;model_family=dit 走 DiT 扩展节(references/dit/evaluate_workflow.md,vbench.py 批量推理产出视频)。Stars: 0Updated Claude CodeCodexUnverifiedInstall
quant-tuning-history-append-ditAscend/msagentAppend a DiT-tuning round to history.yaml. DiT-side counterpart of the LLM-path accuracy_append.py; records practice.md5, inference_outputs, fp_baseline_outputs, and the scoring fields (scores / overall_score / loss_vs_baseline / is_satisfied) populated by quant-tuning-score-dit. Idempotent on practice_id.Stars: 0Updated Claude CodeCodexUnverifiedInstall
quant-tuning-infer-ditAscend/msagent在 W8A8 动态量化后,用 MindIE-SD 冒烟测试 DiT 量化权重是否可加载(可选 skill)。与 `quant-tuning-evaluate`(DiT 扩展节)的区别:本 skill 仅做"能跑通 + 权重可用"的轻量验证(4~16 条 prompt,几分钟出结果),不算分;后者产出视频集交 `quant-tuning-score-dit`(AISBench-VBench 16 维度)评分,是调优回路一环。不引入 vLLM-Ascend、不做精度指标评估(FID/CLIP-score 等)。Stars: 0Updated Claude CodeCodexUnverifiedInstall
quant-tuning-quantizeAscend/msagent执行模型量化。通过 msmodelslim quant 依据 Practice YAML 对模型进行量化。Stars: 0Updated Claude CodeCodexUnverifiedInstall
quant-tuning-quantize-ditAscend/msagentStars: 0Updated Claude CodeCodexUnverifiedInstall