# karpathy/nanochat

> A minimal experimental harness for training LLMs on a single GPU node, covering tokenization, pretraining, finetuning, evaluation and inference. It lets you train a GPT-2-capability model for about $48 and chat with it via CLI.

- Magnitude: 6.8 out of 10 — Breakout
- Stars: 59,898 total · +14 stars measured 2026-11-03, 00:30–01:15 UTC, ≈ 335 by evening
- Star trust: star growth looks organic
- Category: Language models · Language: Python · License: MIT · Created: 2025-10-13 · Last push: 2026-09-07
- GitHub: https://github.com/karpathy/nanochat · Page: https://gitnova.dev/en/r/karpathy/nanochat

## Useful for

- Train your own GPT-2-capability model on 8XH100 and chat with it via CLI
- Set model complexity with a single --depth flag and get auto-tuned hyperparameters
- Reproduce the GPT-2 speedrun and compare training time against the leaderboard

## Why it’s here

- Star-counter measurements on 2026-11-03 (UTC), 00:30–01:15: 59884 → 59898 stars (+14). This is the change over that interval.
- Estimated end-of-day forecast: about +335 stars, using observed gains and the previous day.
- Over the last two days the pace is 8.7× that of the previous week and a half.
- Elevated gains were measured on the last 3 completed UTC days in a row. Continued growth is not guaranteed.
- GitHub Trending today: #14, +321 stars.
- GitHub Trending Python today: #5, +321 stars.
- About 61 forks a day — people are taking the code.

## Star trust

Star growth looks organic.

Star-trust labels are heuristics based on the repository’s behavior, not a check of every stargazer.

## Numbers

- Forks: 8,414
- Issues and pull requests: 708
- Watchers: 378
- Average over the last week: 172 per day
- Usual pace: 32 per day
- Stars in the last hour (measured): 20

## Stars per day, last 30 days (oldest → newest, today is partial)

2026-10-05 … 2026-11-03: 19, 16, 24, 36, 30, 25, 26, 26, 32, 31, 33, 31, 23, 37, 27, 49, 22, 26, 26, 28, 26, 68, 65, 31, 23, 31, 85, 352, 349, 14

## Magnitude by day, last 2 days

11-02 6.9, 11-03 6.8

## Spotted in now

- GitHub Trending today: #14, +321 stars
- GitHub Trending Python today: #5, +321 stars

## Similar by description

1. **KohakuBlueleaf/KohakuMossTrainer** — 0.6 · Steady · Language models · Python · +0 stars measured 2026-11-03, 00:41–01:26 UTC, ≈ 2 by evening
   A finetuning framework for MOSS-Image: every LyCORIS algorithm and full finetune on text-image data at native resolutions, plus inference, on Windows and Linux from 6 GB cards to multi-GPU nodes.
   Full card: https://gitnova.dev/en/r/KohakuBlueleaf/KohakuMossTrainer.md

---

Magnitude (0–10) measures how fast and how unusually interest in a repository is growing right now. It is not a quality score.
Days are UTC. “So far today” is a fact; “expected by the end of the day” is a forecast.
Summaries and use cases are written by an LLM (DeepSeek V4.1 Flash) from the README and may be inaccurate: verify specific claims (benchmarks, speed, hardware) in the repository itself.
Data as of 2026-11-03 01:27 UTC, updated every 30 minutes.