New to Claude Skills? Learn how to install them →

davila7 on GitHub

nanoGPT

Free

Minimalist implementation of GPT for education and experimentation.

Get this skill

Free · Opens the source repo

What nanoGPT does

nanoGPT is a streamlined implementation of the GPT architecture, designed specifically for educational purposes and experimentation with transformer models. With approximately 300 lines of code, it offers a clear and accessible way to understand the inner workings of GPT-2, making it an ideal resource for learners and educators alike. The implementation allows users to train models on character-level data, such as Shakespeare's works, or on larger datasets like OpenWebText, providing flexibility in experimentation.

The skill is built using Python and relies on popular libraries such as PyTorch, NumPy, and Transformers, ensuring compatibility with existing workflows. Users can quickly install the necessary dependencies and follow straightforward commands to prepare data, train models, and generate text. The training process is optimized for both CPU and GPU environments, allowing for rapid experimentation even on limited hardware. For instance, training on Shakespeare can be completed in about five minutes on a CPU, while multi-GPU training of the GPT-2 model can be accomplished in just a few days.

nanoGPT is particularly suitable for those looking to dive deep into the mechanics of transformer models without the overhead of more complex frameworks. It is hackable and encourages modifications, making it a valuable tool for developers and researchers who wish to explore variations of the architecture or training process. The documentation includes detailed workflows and configuration examples, guiding users through the entire process from data preparation to model training and text generation.

This skill is perfect for students, educators, and developers interested in machine learning and natural language processing. It serves as a practical resource for understanding the principles of GPT architecture and experimenting with custom datasets, making it a unique addition to any AI toolkit.

When to use it

Use nanoGPT when you want to learn about GPT architecture, experiment with transformers, or prototype quickly on limited hardware.

When not to use it

This skill is not suitable for production-level applications or large-scale model training, where more robust frameworks like HuggingFace Transformers or Megatron-LM are preferred.

What you can build with it

Learning GPT Architecture

Use nanoGPT to gain a foundational understanding of how GPT models are structured and trained.

Experimenting with Transformers

Quickly prototype different configurations and training setups to see how they affect model performance.

Training on Custom Data

Easily prepare and train models on your own text datasets to explore unique applications of GPT.

How to install nanoGPT

View source

1. Install with the skills CLI

npx skills add davila7/claude-code-templates/model-architecture-nanogpt --agent claude-code

2. Or install it manually

Download the skill folder and drop it into ~/.claude/skills/ for all projects, or .claude/skills/ to scope it to one repo. Restart Claude Code so it picks up the new skill.

Anthropic's agentic coding CLI, and the reference implementation of Agent Skills. Drop a skill folder into ~/.claude/skills and Claude Code loads it automatically whenever a task matches the skill's description. Claude Code docs

Inside SKILL.md

Written by davila7

nanoGPT - Minimalist GPT Training

Quick start

nanoGPT is a simplified GPT implementation designed for learning and experimentation.

Installation:

pip install torch numpy transformers datasets tiktoken wandb tqdm

Train on Shakespeare (CPU-friendly):

# Prepare data
python data/shakespeare_char/prepare.py

# Train (5 minutes on CPU)
python train.py config/train_shakespeare_char.py

# Generate text
python sample.py --out_dir=out-shakespeare-char

Output:

ROMEO:
What say'st thou? Shall I speak, and be a man?

JULIET:
I am afeard, and yet I'll speak; for thou art
One that hath been a man, and yet I know not
What thou art.

Common workflows

Workflow 1: Character-level Shakespeare

Complete training pipeline:

# Step 1: Prepare data (creates train.bin, val.bin)
python data/shakespeare_char/prepare.py

# Step 2: Train small model
python train.py config/train_shakespeare_char.py

# Step 3: Generate text
python sample.py --out_dir=out-shakespeare-char

Config (config/train_shakespeare_char.py):

# Model config
n_layer = 6          # 6 transformer layers
n_head = 6           # 6 attention heads
n_embd = 384         # 384-dim embeddings
block_size = 256     # 256 char context

# Training config
batch_size = 64
learning_rate = 1e-3
max_iters = 5000
eval_interval = 500

# Hardware
device = 'cpu'  # Or 'cuda'
compile = False # Set True for PyTorch 2.0

Training time: ~5 minutes (CPU), ~1 minute (GPU)

Workflow 2: Reproduce GPT-2 (124M)

Multi-GPU training on OpenWebText:

# Step 1: Prepare OpenWebText (takes ~1 hour)
python data/openwebtext/prepare.py

# Step 2: Train GPT-2 124M with DDP (8 GPUs)
torchrun --standalone --nproc_per_node=8 \
  train.py config/train_gpt2.py

# Step 3: Sample from trained model
python sample.py --out_dir=out

Config (config/train_gpt2.py):

# GPT-2 (124M) architecture
n_layer = 12
n_head = 12
n_embd = 768
block_size = 1024
dropout = 0.0

# Training
batch_size = 12
gradient_accumulation_steps = 5 * 8  # Total batch ~0.5M tokens
learning_rate = 6e-4
max_iters = 600000
lr_decay_iters = 600000

# System
compile = True  # PyTorch 2.0

Training time: ~4 days (8× A100)

Workflow 3: Fine-tune pretrained GPT-2

Start from OpenAI checkpoint:

# In train.py or config
init_from = 'gpt2'  # Options: gpt2, gpt2-medium, gpt2-large, gpt2-xl

# Model loads OpenAI weights automatically
python train.py config/finetune_shakespeare.py

Example config (config/finetune_shakespeare.py):

# Start from GPT-2
init_from = 'gpt2'

# Dataset
dataset = 'shakespeare_char'
batch_size = 1
block_size = 1024

# Fine-tuning
learning_rate = 3e-5  # Lower LR for fine-tuning
max_iters = 2000
warmup_iters = 100

# Regularization
weight_decay = 1e-1

Workflow 4: Custom dataset

Train on your own text:

# data/custom/prepare.py
import numpy as np

# Load your data
with open('my_data.txt', 'r') as f:
    text = f.read()

# Create character mappings
chars = sorted(list(set(text)))
stoi = {ch: i for i, ch in enumerate(chars)}
itos = {i: ch for i, ch in enumerate(chars)}

# Tokenize
data = np.array([stoi[ch] for ch in text], dtype=np.uint16)

# Split train/val
n = len(data)
train_data = data[:int(n*0.9)]
val_data = data[int(n*0.9):]

# Save
train_data.tofile('data/custom/train.bin')
val_data.tofile('data/custom/val.bin')

Train:

python data/custom/prepare.py
python train.py --dataset=custom

When to use vs alternatives

Use nanoGPT when:

  • Learning how GPT works
  • Experimenting with transformer variants
  • Teaching/education purposes
  • Quick prototyping
  • Limited compute (can run on CPU)

Simplicity advantages:

  • ~300 lines: Entire model in model.py
  • ~300 lines: Training loop in train.py
  • Hackable: Easy to modify
  • No abstractions: Pure PyTorch

Use alternatives instead:

  • HuggingFace Transformers: Production use, many models
  • Megatron-LM: Large-scale distributed training
  • LitGPT: More architectures, production-ready
  • PyTorch Lightning: Need high-level framework

Common issues

Issue: CUDA out of memory

Reduce batch size or context length:

batch_size = 1  # Reduce from 12
block_size = 512  # Reduce from 1024
gradient_accumulation_steps = 40  # Increase to maintain effective batch

Issue: Training too slow

Enable compilation (PyTorch 2.0+):

compile = True  # 2× speedup

Use mixed precision:

dtype = 'bfloat16'  # Or 'float16'

Issue: Poor generation quality

Train longer:

max_iters = 10000  # Increase from 5000

Lower temperature:

# In sample.py
temperature = 0.7  # Lower from 1.0
top_k = 200       # Add top-k sampling

Issue: Can't load GPT-2 weights

Install transformers:

pip install transformers

Check model name:

init_from = 'gpt2'  # Valid: gpt2, gpt2-medium, gpt2-large, gpt2-xl

Advanced topics

Model architecture: See references/architecture.md for GPT block structure, multi-head attention, and MLP layers explained simply.

Training loop: See references/training.md for learning rate schedule, gradient accumulation, and distributed data parallel setup.

Data preparation: See references/data.md for tokenization strategies (character-level vs BPE) and binary format details.

Hardware requirements

  • Shakespeare (char-level):

    • CPU: 5 minutes
    • GPU (T4): 1 minute
    • VRAM: <1GB
  • GPT-2 (124M):

    • 1× A100: ~1 week
    • 8× A100: ~4 days
    • VRAM: ~16GB per GPU
  • GPT-2 Medium (350M):

    • 8× A100: ~2 weeks
    • VRAM: ~40GB per GPU

Performance:

  • With compile=True: 2× speedup
  • With dtype=bfloat16: 50% memory reduction

Resources

Frequently asked questions about nanoGPT

Similar skills