Tilegym Cutile Python
nvidia/skills/skills/tilegym-cutile-python作者 nvidiacf5224d14250CC-BY-4.0 AND Apache-2.03.5K 个星标收录于 2026年10月8日更新于 2026年10月8日仓库今天更新
Expert cuTile programming assistant. Write high-performance GPU kernels using cuTile's tile-based programming model with proper validation and optimization. Supports deep agent orchestration for complex multi-kernel tasks.
仅公开文件列表。将技能安装到工作区后即可查看文件内容。
| 路径 | 大小 | 类型 |
|---|---|---|
| BENCHMARK.md | 7.2 KB | text/markdown |
| evals/evals.json | 3.9 KB | application/json |
| examples/convolution/conv2d_with_bias_dilation_groups.py | 14.5 KB | text/plain |
| examples/convolution/conv3d_with_bias_dilation_groups.py | 15.3 KB | text/plain |
| examples/convolution/conv_transpose_2d.py | 14.3 KB | text/plain |
| examples/convolution/conv_transpose_3d.py | 15.6 KB | text/plain |
| examples/convolution/README.md | 2.9 KB | text/markdown |
| examples/matmul/matmul_4d_tensors.py | 4.2 KB | text/plain |
| examples/matmul/matrix_vector_multiplication.py | 2.4 KB | text/plain |
| examples/matmul/README.md | 429 B | text/markdown |
| examples/matmul/split_k_gemm.py | 4.8 KB | text/plain |
| examples/normalization/group_norm.py | 5.5 KB | text/plain |
| examples/normalization/README.md | 241 B | text/markdown |
| examples/pooling/avgpool3d.py | 6.6 KB | text/plain |
| examples/pooling/maxpool3d.py | 7.3 KB | text/plain |
| examples/pooling/README.md | 121 B | text/markdown |
| examples/scan/cumsum_cumprod_blocking.py | 2.1 KB | text/plain |
| examples/scan/README.md | 116 B | text/markdown |
| examples/tilegym_and_examples_guide.md | 3.3 KB | text/markdown |
| guidelines/01_implementation_lessons.md | 8.1 KB | text/markdown |
| guidelines/02_code_generation_rules.md | 10.6 KB | text/markdown |
| guidelines/03_concepts.md | 4.7 KB | text/markdown |
| orchestration/analyzer_agent.md | 11.7 KB | text/markdown |
| orchestration/composer_agent.md | 11.2 KB | text/markdown |
| orchestration/kernel_agent.md | 7.3 KB | text/markdown |
| orchestration/overview.md | 9.7 KB | text/markdown |
| orchestration/workflow.md | 8.9 KB | text/markdown |
| skill-card.md | 4 KB | text/markdown |
| SKILL.md | 19.9 KB | text/markdown |
| skill.oms.sig | 12.4 KB | text/plain |
| torch-learner/examples/lstm_trace.md | 11.2 KB | text/markdown |
| torch-learner/references/1_pytorch_codebase_map.md | 8.9 KB | text/markdown |
| torch-learner/references/2_dispatch_mechanism.md | 8.9 KB | text/markdown |
| torch-learner/references/3_tracing_strategies.md | 8.8 KB | text/markdown |
| torch-learner/references/4_language_layers.md | 12.7 KB | text/markdown |
| torch-learner/references/5_well_known_ops.md | 11.2 KB | text/markdown |
| torch-learner/tracing_workflow.md | 7.2 KB | text/markdown |
来源与署名
来源：nvidia/skills位于skills/tilegym-cutile-python提交cf5224d
许可证： CC-BY-4.0 AND Apache-2.0
内容归原作者所有。SourceWeft 从公开仓库中收录这些内容。
更多来自 nvidia/skills 的技能
Tilegym Monkey Patch Kernels To Transformers
nvidia
通过非侵入式 monkey-patch 将 TileGym cuTile 内核集成到 Hugging Face transformers 模型，并自动生成新内核。
Tilegym Converting Cutile To Julia
nvidia
将 cuTile Python GPU 内核转换为 cuTile.jl Julia 等价实现，并提供规则、示例与静态校验器。
Tilegym Adding Cutile Kernel
nvidia
指导向 TileGym 添加新的 cuTile GPU 内核算子，涵盖调度注册、测试与基准测试。
Physicsnemo Discover
nvidia
为 SciML 任务指引查找 PhysicsNeMo 的模型族、数据管道与示例，并引用实时仓库路径。
Nemoclaw User Guide
nvidia
引导 AI 代理访问 NemoClaw 官方 Markdown 文档与文档 MCP 服务器，回答安装与运维问题。
Nemo Rl Docs
nvidia
NeMo-RL 的文档规范：维护 docs/index.md 并使用 Google 风格文档字符串。
更多Software Development技能
Playground
anthropics
构建自包含的交互式 HTML 演练场，包含控件、实时预览和可复制的提示词输出。
M5 Onboard
anthropics
通过 USB 检测 M5Stack ESP32 开发板，刷写 UIFlow 2.0 固件并安装 MicroPython 应用包。
Cardputer Buddy
anthropics
指导迭代 Cardputer-Adv 的 MicroPython 应用包：添加应用、通过 USB 串口推送文件、查看日志并运行 REPL 命令。
Web Design Guidelines
vercel-labs
依据从远程来源获取的 Web Interface Guidelines 审查 UI 代码，并以 file:line 形式报告问题。
Vercel React Native Skills
vercel-labs
React Native 与 Expo 最佳实践规则，涵盖列表性能、动画、导航、UI 模式、状态与 monorepo 配置。
Vercel React Best Practices
vercel-labs
Vercel 的 React 与 Next.js 性能指南，以 70 条规则的形式用于编写、审查和重构代码。