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.

PyTorch training on several GPUs with Accelerate v1.0.0

Use this resource to adapt a PyTorch script to several GPUs or nodes with minimal changes

New

  • 570 downloads
  • 172 likes
  • 123 saves

Free

Install with the CLI

In a terminal, from the folder you want:

npx --yes palatekit@latest install multi-gpu-training-accelerate

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 Accelerate library to a PyTorch script, prepares the model, optimiser and data, then sets up launching on one GPU, several GPUs or several nodes, with mixed precision if needed. You get a distributed training script.

Mostly picked by

Updated November 2, 2026

Card for the “PyTorch training on several GPUs with Accelerate” skill

Details

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

Similar resources

View all

What you get

  • SKILL.md: moving from one GPU to many, multi-node training and mixed precision
  • references/: custom plugins, Megatron integration and optimisations

Good fit for

  • Your training runs on one GPU and should use eight
  • You want FP16 or BF16 without rewriting the training loop

Not for

    Tags

    • pytorch
    • gpu
    • distributed
    • accelerate
    • training

    About

    You say

    “Adapt my training script for eight GPUs and enable BF16.”

    Card for the “PyTorch training on several GPUs with Accelerate” 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 adapt a PyTorch script to several GPUs or nodes with minimal changes
    3. Review and adjust the result

    Content preview

    Files

    • LICENSE1 KB
    • NOTICE-PALATEKIT.md0.5 KB
    • references/custom-plugins.md11.5 KB
    • references/megatron-integration.md11 KB
    • references/performance.md12.3 KB
    • SKILL.md9 KB

    Origin and licence

    Listed unmodified

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

    Notice

    « accelerate » (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.