Every premium and members-only resource for $9.99/month, as a no-commitment subscription. See the plans

PalateKit

Sign in
Menu

Request a resource

Describe the resource you are missing in one sentence. We will email you when it is published.

0 / 280

A page, a video or an image online.

Unlock PalateKit

Unlock everything

Plan

Paying means accepting our terms of sale.

Secure payment by Stripe. Receipts and invoices sent by email.

Welcome back

Sign in to install, download or save this resource. A free account is enough for free resources.

Fast, memory-efficient attention with FlashAttention v1.0.0

Use this resource to speed up transformer attention on long sequences

New

  • 430 downloads
  • 200 likes
  • 215 saves

Free

Install with the CLI

In a terminal, from the folder you want:

npx --yes palatekit@latest install fast-attention-flash-attention

With no option, the skill goes to ~/.agents/skills/; add --agent claude-code (or codex, cursor…) to target an agent.

First time: connect the CLI with

npx --yes palatekit@latest login

The agent turns on Flash attention in an existing PyTorch model, through PyTorch 2.2's native attention or the dedicated package, then compares outputs and performance. You get a faster, more memory-efficient model.

Mostly picked by

Updated November 2, 2026

Card for the “Fast, memory-efficient attention with FlashAttention” skill

Details

Works in
Install guides
Generates with
Your agent's model
Licence
MIT

Similar resources

View all

What you get

  • SKILL.md: quick start, enabling it in an existing model and comparison
  • references/: benchmarks and integration with Transformers

Good fit for

  • Your long sequences exhaust GPU memory
  • You want more speed without changing the architecture

Not for

    Tags

    • attention
    • transformers
    • gpu
    • pytorch
    • performance

    About

    You say

    “Turn on FlashAttention in my model and check that the outputs stay equivalent.”

    Card for the “Fast, memory-efficient attention with FlashAttention” skill

    Skill published by Orchestra Research under the MIT license, delivered in its original version.

    Show more

    How to use it

    1. Install the skill in your agent's skills folder
    2. Ask your agent to speed up transformer attention on long sequences
    3. Review and adjust the result

    Content preview

    Files

    • LICENSE1 KB
    • NOTICE-PALATEKIT.md0.5 KB
    • references/benchmarks.md7 KB
    • references/transformers-integration.md7.3 KB
    • SKILL.md10.4 KB

    Origin and licence

    Listed unmodified

    Author
    Orchestra Research
    Source
    github.com/nousresearch/hermes-agent/tree/0c74cc682dc81bf78b78a1737b3ee77a820ccb79/optional-skills/mlops/flash-attention
    Licence
    MIT

    Notice

    « flash-attention » (Orchestra Research, MIT) livré dans sa version d'origine : fiche en français et en anglais par PalateKit.

    User reviews

    –

    No ratings yet

    Loading reviews…

    Delete your review?

    Your rating and comment will be removed from the page.