nanogpt-training

Community

Single-GPU GPT-2 scale training.

AuthorKaiserWhoLearns
Version1.0.0
Installs0

System Documentation

What problem does it solve?

Train GPT-2 scale models efficiently on a single GPU, enabling rapid prototyping without heavy hardware.

Core Features & Use Cases

  • Efficient single-GPU pretraining for GPT-2-scale architectures (e.g., ~124M parameters).
  • Tokenized data loading from HuggingFace Hub or local shards with memory-friendly handling.
  • Mixed-precision training and configurable optimizers (AdamW, Muon) with a flexible training loop for experimentation.

Quick Start

Run the included minimal training script to reproduce a GPT-2-scale training workflow on a single GPU.

Dependency Matrix

Required Modules

None required

Components

references

💻 Claude Code Installation

Recommended: Let Claude install automatically. Simply copy and paste the text below to Claude Code.

Please help me install this Skill:
Name: nanogpt-training
Download link: https://github.com/KaiserWhoLearns/skillsbench/archive/main.zip#nanogpt-training

Please download this .zip file, extract it, and install it in the .claude/skills/ directory.
View Source Repository

Agent Skills Search Helper

Install a tiny helper to your Agent, search and equip skill from 471,000+ vetted skills library on demand.