Loading...
Loading...
Found 184 Skills
GENERator DNA 序列生成模型的昇腾 NPU 迁移 Skill，适用于将基于 HuggingFace Transformers 的 Causal LM 从 CUDA 迁移到华为 Ascend NPU，覆盖环境搭建、依赖安装、代码适配、多进程处理和 sequence recovery 验证。
将简单Vector类型Triton算子从GPU迁移到昇腾NPU。当用户需要迁移Triton代码到NPU、提到GPU到NPU迁移、Triton迁移、昇腾适配时使用。注意：无法自动迁移存在编译问题的算子。
Guide spec-driven development using collaborative interrogation and iterative Q&A to build production-ready specifications. Use when the user wants to build specifications, plan features, gather requirements, create technical blueprints, or asks about spec-driven development, requirements gathering, feature planning, or specification writing.
Guides PRD, ADR, Design Doc, and Work Plan creation with templates and decision matrix.
Use when testing MCP servers, debugging MCP tool responses, exploring MCP capabilities, or diagnosing why an MCP tool returns unexpected data
Calculate MFU (Machine FLOP Utilization) for operators such as matmul/GEMM, and provide clear formulas and derivation processes.
MindSpeed-LLM 环境搭建指南，用于华为昇腾 NPU。覆盖 CANN 环境激活、PyTorch + torch_npu 安装、MindSpeed 加速库安装、Megatron-LM 核心模块集成、MindSpeed-LLM 安装及环境验证。当用户需要在昇腾 NPU 上搭建 MindSpeed-LLM 训练环境时使用。
Used for reviewing GitCode PRs, generating in-depth review conclusions or publishing line-by-line comments by combining PR metadata, diffs, and the context of the entire code repository. It is used when users want to review a GitCode PR, check a GitCode PR link, analyze change risks, or publish review comments to a GitCode PR. Typical trigger phrases include "review this PR", "inspect this PR", "check PR", or directly providing a GitCode PR link, such as https://gitcode.com/owner/repo/pull/123.
Create detailed feature specifications with user stories, acceptance criteria, and edge cases. Use when starting a new feature or initializing a new project.
Privy SDK Documentation
DeepFRI 的 TensorFlow 到 PyTorch 转换与昇腾 NPU 迁移 Skill，适用于蛋白质功能预测场景下的 TF 模型分析、PyTorch 重写、权重逐层映射、NPU 推理与精度验证，尤其适合需要在 Ascend 上运行 DeepFRI CNN 或 GCN 路径时使用。
Migrate GPU/CUDA Triton operators to Triton-Ascend, or rewrite Python/PyTorch operators into Triton-Ascend implementations that can run on Ascend NPU. When clear optimization opportunities are identified, directly output the optimized code, minimal validation script, and troubleshooting instructions. This skill should be prioritized when users mention 昇腾 (Ascend), Ascend, NPU, triton-ascend, Triton operator migration, PyTorch operator rewriting, coreDim, UB overflow, 1D grid, physical core binding, block_ptr, stride, memory access alignment, mask performance, dtype degradation, operator optimization, or directly ask questions like "How to use this skill", "How to run it in the command line", "How to perform migration/validation in a container", even if users do not explicitly say "write a skill" or "perform migration".