Nanogpt
/SKILLImplémentation pédagogique de GPT en environ 300 lignes de code. Reproduit les performances de GPT-2 (124M) sur OpenWebText. Code épuré et modulable, idéal pour l'apprentissage des modèles
name: nanogpt
description: Educational GPT implementation in ~300 lines. Reproduces GPT-2 (124M) on OpenWebText. Clean, hackable code for learning transformers. By Andrej Karpathy. Perfect for understanding GPT architecture from scratch. Train on Shakespeare (CPU) or OpenWebText (multi-GPU).
version: 1.0.0
author: Orchestra Research
license: MIT
tags: [Model Architecture, NanoGPT, GPT-2, Educational, Andrej Karpathy, Transformer, Minimalist, From Scratch, Training]
dependencies: [torch, transformers, datasets, tiktoken, wandb]
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 tqdmTrain 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-charOutput:
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-charConfig (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 2.0Training 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=outConfig (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 2.0Training 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.pyExample 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-1Workflow 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=customWhen 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 batchIssue: Training too slow
Enable co
