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

PalateKit

Menu

Request a resource

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

0 / 280

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.

Choosing the MoE token dispatcher v1.0.0

Use this resource to choose the MoE token dispatcher that fits your hardware and EP degree

New

  • 381 downloads
  • 106 likes
  • 239 saves

Free

Install with the CLI

In a terminal, from the folder you want:

npx --yes palatekit@latest install moe-token-dispatcher-selection

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 recommends a dispatcher from the hardware (H100, B200, GB200 or GB300), the expert-parallel degree and the results seen on DSV3 and Qwen3, then names the settings to tune and backend availability. You get a justified choice and the parameters to adjust.

Mostly picked by

Updated November 2, 2026

Card for the “Choosing the MoE token dispatcher” skill

Details

Generates with
Your agent's model
Licence
Apache-2.0

Similar resources

View all

What you get

  • SKILL.md: selection guide and rules
  • card.yaml: structured card
  • evals/evals.json: evaluation cases

Good fit for

  • You start a MoE on GB200 and hesitate between backends
  • Your alltoall dispatch is too slow at a large EP degree

Not for

    Tags

    • MoE
    • dispatcher
    • DeepEP
    • HybridEP

    About

    You say

    “Which MoE dispatcher should I use for Qwen3 235B on GB200?”

    Card for the “Choosing the MoE token dispatcher” 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 choose the MoE token dispatcher that fits your hardware and EP degree
    3. Review and adjust the result

    Content preview

    Files

    • BENCHMARK.md4.4 KB
    • card.yaml4.1 KB
    • evals/evals.json1.5 KB
    • LICENSE11.3 KB
    • NOTICE-PALATEKIT.md0.5 KB
    • skill-card.md4 KB
    • SKILL.md8 KB
    • skill.oms.sig4.7 KB

    Origin and licence

    Listed unmodified

    Author
    nvidia
    Source
    github.com/nvidia/skills/tree/ca116d60beafe1b4459f0ab790b7b450912c56b0/skills/nemo-mbridge-perf-moe-dispatcher-selection
    Licence
    Apache-2.0

    Notice

    « nemo-mbridge-perf-moe-dispatcher-selection » (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.