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 a small GPT step by step with nanoGPT v1.0.0

Use this resource to understand GPT architecture by training a small model on CPU

New

  • 403 downloads
  • 109 likes
  • 130 saves

Free

Install with the CLI

In a terminal, from the folder you want:

npx --yes palatekit@latest install nanogpt-minimal-gpt-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 preparing data, training a small model on Shakespeare in a few minutes on CPU, then generating text. It relies on a GPT implementation of about 300 lines, meant to be read and modified.

Mostly picked by

Updated November 2, 2026

Card for the “Training a small GPT step by step with nanoGPT” skill

Details

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

Similar resources

View all

What you get

  • SKILL.md: data preparation, training and generation
  • references/architecture.md, data.md and training.md: architecture, data and training settings

Good fit for

  • You want to understand GPT by reading the code
  • You want a first training run without a GPU

Not for

    Tags

    • gpt
    • teaching
    • transformer
    • training

    About

    You say

    “Train a small GPT on Shakespeare on my computer and explain each step.”

    Card for the “Training a small GPT step by step with nanoGPT” 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 understand GPT architecture by training a small model on CPU
    3. Review and adjust the result

    Content preview

    Files

    • LICENSE1.1 KB
    • NOTICE-PALATEKIT.md0.5 KB
    • references/architecture.md11.8 KB
    • references/data.md11.4 KB
    • references/training.md13.8 KB
    • SKILL.md6.9 KB

    Origin and licence

    Listed unmodified

    Author
    Orchestra Research
    Source
    github.com/orchestra-research/ai-research-skills/tree/773a52944ba4747a18bd4ae9ade53fff041adcbc/01-model-architecture/nanogpt
    Licence
    MIT

    Notice

    « nanogpt » (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.