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.

Training large LLMs with Megatron-Core v1.0.0

Use this resource to train multi-billion-parameter LLMs with 3D parallelism

New

  • 312 downloads
  • 171 likes
  • 229 saves

Free

Install with the CLI

In a terminal, from the folder you want:

npx --yes palatekit@latest install megatron-core-large-llm-training-2

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 guides you through installing Megatron-Core (Docker or pip), running a simple two-GPU training, then configuring a LLaMA-style model with tensor, pipeline and data parallelism. It covers training settings and an eight-GPU setup.

Mostly picked by

Updated November 2, 2026

Card for the “Training large LLMs with Megatron-Core” skill

Details

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

Similar resources

View all

What you get

  • SKILL.md: installation, simple training and LLaMA example
  • references/parallelism-guide.md, training-recipes.md, benchmarks.md and production-examples.md: parallelism, recipes, benchmarks and production examples

Good fit for

  • You train models of more than one billion parameters
  • You want tensor and pipeline parallelism

Not for

    Tags

    • training
    • parallelism
    • megatron
    • nvidia

    About

    You say

    “Set up a LLaMA 8B-style training with 3D parallelism on one 8-GPU node.”

    Card for the “Training large LLMs with Megatron-Core” 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 train multi-billion-parameter LLMs with 3D parallelism
    3. Review and adjust the result

    Content preview

    Files

    • LICENSE1.1 KB
    • NOTICE-PALATEKIT.md0.5 KB
    • references/benchmarks.md7.4 KB
    • references/parallelism-guide.md9.8 KB
    • references/production-examples.md11.1 KB
    • references/training-recipes.md11.9 KB
    • SKILL.md9.8 KB

    Origin and licence

    Listed unmodified

    Author
    Orchestra Research
    Source
    github.com/orchestra-research/ai-research-skills/tree/773a52944ba4747a18bd4ae9ade53fff041adcbc/08-distributed-training/megatron-core
    Licence
    MIT

    Notice

    « training-llms-megatron » (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.