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.

Large-scale LLM pretraining with TorchTitan v1.0.0

Use this resource to pretrain a large language model with PyTorch's 4D parallelism

New

  • 475 downloads
  • 235 likes
  • 206 saves

Free

Install with the CLI

In a terminal, from the folder you want:

npx --yes palatekit@latest install llm-pretraining-torchtitan-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 installs TorchTitan, downloads the tokenizer, picks a configuration by name from the registry, then runs pretraining on one or several nodes. You get a distributed pretraining run, with the fields you overrode.

Mostly picked by

Updated November 2, 2026

Card for the “Large-scale LLM pretraining with TorchTitan” skill

Details

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

Similar resources

View all

What you get

  • SKILL.md: installation, tokenizer download and training procedure
  • references/: checkpoints, custom models, float8 and FSDP

Good fit for

  • You pretrain a Llama-style model on several GPUs
  • You want to compare parallelism settings on the same model

Not for

    Tags

    • pretraining
    • pytorch
    • fsdp
    • distributed
    • llm

    About

    You say

    “Launch Llama 3.1 8B pretraining on one node of eight GPUs.”

    Card for the “Large-scale LLM pretraining with TorchTitan” 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 pretrain a large language model with PyTorch's 4D parallelism
    3. Review and adjust the result

    Content preview

    Files

    • LICENSE1 KB
    • NOTICE-PALATEKIT.md0.5 KB
    • references/checkpoint.md4.1 KB
    • references/custom-models.md7.2 KB
    • references/float8.md4 KB
    • references/fsdp.md3.8 KB
    • SKILL.md10.1 KB

    Origin and licence

    Listed unmodified

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

    Notice

    « torchtitan » (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.