# jingyaogong/minimind

> An open-source educational project for training a tiny 64M-parameter LLM (MiniMind) from scratch in PyTorch, covering the full pipeline: pretrain, SFT, LoRA, RLHF, RLAIF and Tool Use.

- Magnitude: 2.8 out of 10 — Steady
- Stars: 63,594 total · +34 stars measured 2026-10-12, 00:15–11:01 UTC, ≈ 74 by evening
- Star trust: star growth looks organic
- Category: Language models · Language: Python · License: Apache-2.0 · Created: 2024-07-27 · Last push: 2026-09-22
- GitHub: https://github.com/jingyaogong/minimind · Homepage: https://jingyaogong.github.io/minimind · Page: https://gitnova.dev/en/r/jingyaogong/minimind

## Useful for

- Train your own tiny LLM from scratch on a single GPU in a couple of hours
- Study pretrain, SFT, LoRA and RLHF implementations without transformers/trl wrappers
- Deploy an OpenAI-API-compatible MiniMind server for chat UIs

## Why it’s here

- Star-counter measurements on 2026-10-12 (UTC), 00:15–11:01: 63560 → 63594 stars (+34). This is the change over that interval.
- Estimated end-of-day forecast: about +74 stars, using observed gains and the previous day.
- About 7 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: 8,259
- Issues and pull requests: 801
- Watchers: 286
- Average over the last week: 68 per day
- Usual pace: 107 per day
- Latest release: v2 (2025-10-21)

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

2026-10-04 … 2026-11-02: 46, 39, 37, 55, 95, 87, 52, 76, 10, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None

## Magnitude by day, last 14 days

09-29 2.6, 09-30 2.6, 10-01 2.6, 10-02 2.5, 10-03 2.3, 10-04 2.2, 10-05 2.3, 10-06 2.3, 10-07 2.3, 10-08 3.0, 10-09 3.4, 10-10 3.3, 10-11 2.6, 10-12 2.8

---

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 2026-10-12 11:12 UTC, updated every 30 minutes.