# marin-community/marin

> Open platform and community for research and development of foundation models: data curation, tokenization, pretraining, posttraining and evaluation of LLMs. Aimed at researchers and engineers training language and multimodal models.

- Magnitude: 1.1 out of 10 — Steady
- Stars: 3,833 total · +1 star on 2026-09-29 (UTC), ≈ 3 by evening
- Star trust: star growth looks organic
- Category: Language models · Language: Python · License: Apache-2.0 · Created: 2024-03-22 · Last push: 2026-09-29
- GitHub: https://github.com/marin-community/marin · Homepage: https://marin.community · Page: https://gitnova.dev/en/r/marin-community/marin

## Useful for

- Train a tiny language model on TinyStories using the README example
- Build a tokenization-and-training pipeline as a set of dependent steps
- Reproduce training mixtures from Nemotron-CC, StarCoderData and ProofPile 2

## Why it’s here

- 1 star so far today, about 3 expected by the end of the day.

## 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: 309
- Issues and pull requests: 9,533
- Watchers: 46
- Average over the last week: 6 per day
- Usual pace: 31 per day
- Latest release: dev-wheels (2026-08-13)

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

2026-10-05 … 2026-11-03: None, None, None, None, None, None, None, None, None, 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 12 days

09-18 1.8, 09-19 1.4, 09-20 1.5, 09-21 1.5, 09-22 1.3, 09-23 1.4, 09-24 1.3, 09-25 1.2, 09-26 0.9, 09-27 1.1, 09-28 1.2, 09-29 1.2

---

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 2026-09-29 10:48 UTC, updated every 30 minutes.