nanochat

by Andrej Karpathy

Read the original →

A compact training harness spanning the model-development lifecycle.

Karpathy's nanochat connects tokenization, pretraining, fine-tuning, evaluation and inference in one repository. It is intended for studying and modifying a real training stack, not merely trying a chat interface.

Level
Expert
Models
any
Category
Model Training
Tags
Source date
date not stated
Added
2026-10-18
Last verified
2026-10-18
Status
ok
Prerequisites
Acts on model-training, infrastructure or research tooling; requires training-stack expertise.
Subcategory
pretraining
Region
West
Original language
en
Access
free
Resource reuse
open: Source states MIT. Applies to the linked resource, not every referenced model or dataset. Facet definitions

Summary in our own words. Rights remain with the original creator. Credit and correction policy.

Plain-text Markdown copy