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.

Speculative decoding on a Jetson vLLM server v1.0.0

Use this resource to speed up token generation on a Jetson vLLM server with speculative decoding

New

  • 562 downloads
  • 171 likes
  • 184 saves

Free

Install with the CLI

In a terminal, from the folder you want:

npx --yes palatekit@latest install vllm-speculative-decoding-jetson

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 adds the speculative decoding configuration to your existing server, accounts for the Jetson-specific GPU memory headroom, then measures whether decode throughput improves. It also notes when the technique does not help.

Mostly picked by

Updated November 2, 2026

Card for the “Speculative decoding on a Jetson vLLM server” skill

Details

Works in
Install guides
Generates with
Your agent's model
Licence
Apache-2.0

Similar resources

View all

What you get

  • SKILL.md: Jetson tuning rules, verification and limits
  • evals/evals.json: test cases
  • BENCHMARK.md: evaluation criteria

Good fit for

  • Your vLLM server on Jetson answers well but generates slowly
  • You want to test EAGLE-3 on your model without changing hardware

Not for

    Tags

    • jetson
    • vllm
    • speculative decoding
    • llm

    About

    You say

    “Add EAGLE-3 speculative decoding to my Jetson vLLM server and measure the gain.”

    Card for the “Speculative decoding on a Jetson vLLM server” skill

    Skill published by nvidia under the Apache-2.0 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 token generation on a Jetson vLLM server with speculative decoding
    3. Review and adjust the result

    Content preview

    Files

    • BENCHMARK.md3.4 KB
    • evals/evals.json6.4 KB
    • LICENSE11.3 KB
    • NOTICE-PALATEKIT.md0.5 KB
    • skill-card.md3.7 KB
    • SKILL.md4.8 KB
    • skill.oms.sig4.5 KB

    Origin and licence

    Listed unmodified

    Author
    nvidia
    Source
    github.com/nvidia/skills/tree/ca116d60beafe1b4459f0ab790b7b450912c56b0/skills/jetson-speculative-decoding
    Licence
    Apache-2.0

    Notice

    « jetson-speculative-decoding » (nvidia, Apache-2.0) 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.