skills.rest
fla-org
fla-orgOFFICIAL
Explore all skills by this author →
GitHub

fla-kda

Manage KDA development tasks within the Flash Linear Attention framework.

5.5k|643|Updated Dec 20, 2023
Search Tags:#software development#code optimization#KDA#kernel workflow#Flash Linear Attention
One-click install
npx skills add https://github.com/fla-org/flash-linear-attention --skill fla-kda
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: fla-kda
Source: https://github.com/fla-org/flash-linear-attention/tree/main/.agents/skills/fla-kda
Command: npx skills add https://github.com/fla-org/flash-linear-attention --skill fla-kda

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) components.

What problem does it solve?

The fla-kda Skill streamlines the modification, review, and testing of KDA components within the Flash Linear Attention framework, enhancing workflow efficiency and code quality.

Core Features & Use Cases

  • KDA Workflow Management: Provides a structured environment for KDA-specific tasks.
  • Public Technical Notes: Offers detailed information on KDA gates, chunk kernels, and backends.
  • Gate Modes: Supports both pre-gated and in-kernel gate modes with safe and non-safe options.
  • Use Case: When working on KDA-related code under fla/ops/kda/**, use this skill to ensure consistent and efficient development practices.

Quick Start

Utilize the fla-kda skill to execute KDA-related modifications or reviews within the fla/ops/kda/** directory.

Frequently Asked Questions about fla-kda

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I modify KDA chunk kernels within the Flash Linear Attention framework?▼

The Skill streamlines KDA chunk kernel modifications within the Flash Linear Attention framework by providing a structured environment for code changes, reviews, and testing specifically under the `fla/ops/kda/**` directory.

What are pre-gated and in-kernel gate modes in KDA operations?▼

KDA gate modes include pre-gated and in-kernel options, both supporting safe and non-safe variations. These modes control how gates are processed during KDA operations within the Flash Linear Attention framework.

How do I test and benchmark KDA code in Flash Linear Attention?▼

You can test and benchmark KDA code in Flash Linear Attention by applying this Skill to your development process. It offers structured workflows and public technical notes to guide modifications and performance evaluations for KDA components.

Does the Flash Linear Attention framework provide specialized workflows for KDA backends?▼

Yes, the Flash Linear Attention framework provides specialized workflows for KDA backends. This Skill offers detailed technical notes and structured management for developing, modifying, and reviewing backend components efficiently.

When should I use a specialized workflow for KDA code modifications?▼

You should use a specialized workflow for KDA code modifications when working under the `fla/ops/kda/**` directory, ensuring consistent development practices, proper gate mode configuration, and efficient testing of chunk kernels and backends.

Related Skills

Explore More Software Engineering Skills→
fla-orgfla-org
5.5k

fla-nvidia-performance

Profile and benchmark NVIDIA GPU kernels in the FLA repository.

Official
Advanced
phraphra

code-review

Review modified code files and auto-fix mechanical issues.

Community
Advanced
fla-orgfla-org
5.5k

fla-dispatch-backends

Manage and optimize backend dispatch for Flash Linear Attention operations.

Official
Advanced
projectedanxprojectedanx
1

requesting-code-review

Automate pre-commit code verification with security analysis, linting, and subagent review.

Community
Advanced