Escribe kernels Metal/MPS para operadores de PyTorch. Útil al agregar soporte para dispositivos MPS a operadores, implementar shaders Metal o portar kernels CUDA a Apple Silicon. Cubre el despacho de native_functions.yaml, operadores del lado del host y la implementación del kernel Metal.
Esta habilidad está diseñada para desarrolladores que trabajan con PyTorch en hardware Apple Silicon y necesitan escribir kernels Metal/MPS. Su objetivo principal es facilitar la adición de soporte para dispositivos MPS a operadores de PyTorch, ya sea implementando nuevos shaders Metal o portando kernels CUDA existentes a la arquitectura Apple Silicon. La habilidad cubre tres áreas clave: el despacho de operadores a través del archivo `native_functions.yaml`, la implementación de operadores del lado del host (C++), y la escritura de kernels Metal (shaders) para la GPU. Es útil para cualquier persona que quiera optimizar el rendimiento de sus modelos en Macs con chips M1/M2/M3, asegurando que los operadores personalizados funcionen de manera eficiente en el backend MPS de PyTorch. La documentación asume familiaridad con PyTorch, C++ y el lenguaje de shaders Metal, pero proporciona una guía estructurada para integrar estos componentes.
npx skills add https://github.com/pytorch/pytorch/tree/main/skills/metal-kernelInstalled? Explore more Programación y desarrollo skills: steipete/bluebubbles, steipete/eightctl, steipete/blucli · View all 6 →
Escribe kernels Metal/MPS para operadores de PyTorch. Úsalo al agregar soporte para dispositivos MPS a operadores, implementar shaders Metal o portar kernels CUDA a Apple Silicon. Cubre el despacho de native_functions.yaml, operadores del lado del host y la implementación del kernel Metal.
This repo contains 8 individual skills — each has its own dedicated page.
Add unsigned integer (uint) type support to PyTorch operators by updating AT_DISPATCH macros. Use when adding support for uint16, uint32, uint64 types to operators, kernels, or when user mentions enabling unsigned types, barebones unsigned types, or uint support.
Debug AOTInductor (AOTI) errors and crashes. Use when encountering AOTI segfaults, device mismatch errors, constant loading failures, or runtime errors from aot_compile, aot_load, aoti_compile_and_package, or aoti_load_package.
Convert PyTorch AT_DISPATCH macros to AT_DISPATCH_V2 format in ATen C++ code. Use when porting AT_DISPATCH_ALL_TYPES_AND*, AT_DISPATCH_FLOATING_TYPES*, or other dispatch macros to the new v2 API. For ATen kernel files, CUDA kernels, and native operator implementations.
Write docstrings for PyTorch functions and methods following PyTorch conventions. Use when writing or updating docstrings in PyTorch code.
Review PyTorch pull requests for code quality, test coverage, security, and backward compatibility. Use when reviewing PRs, when asked to review code changes, or when the user mentions "review PR", "code review", or "check this PR".
Migrate a file to use stricter Pyrefly type checking with annotations required for all functions, classes, and attributes.
Guide users through creating Agent Skills for Claude Code. Use when the user wants to create, write, author, or design a new Skill, or needs help with SKILL.md files, frontmatter, or skill structure.
Triages GitHub issues by routing to oncall teams, applying labels, and closing questions. Use when processing new PyTorch issues or when asked to triage an issue.
Send and manage iMessages via BlueBubbles, including attachments, tapbacks, edits, replies, and groups.
Control Eight Sleep pods (status, temperature, alarms, schedules).
BluOS CLI (blu) for discovery, playback, grouping, and volume.
Create, search, and manage Bear notes via grizzly CLI.
Capture frames or clips from RTSP/ONVIF cameras.
Search GIF providers with CLI/TUI, download results, and extract stills/sheets.