# shyamsn97/mario-gpt

[Neurips 2023] Generating Mario Levels with GPT2. Code for the paper "MarioGPT: Open-Ended Text2Level Generation through Large Language Models" https://arxiv.org/abs/2302.05981

Repository: https://github.com/shyamsn97/mario-gpt
Canonical: https://ross.abutalabs.com/products/mario-gpt
Homepage: https://huggingface.co/shyamsn97/Mario-GPT2-700-context-length
Language: Python
License: MIT
License Family: permissive
Last push: 2024-07-22T14:09:16+00:00

## Health v2 (maintenance only)
Score: 21/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 0, release rhythm 8, longevity 92
- inputs: {"age_days": 1296, "days_push": 772, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1152, forks 106 (observed 2026-08-28T04:03:46.892078+00:00)

## What it is
MarioGPT is a Python library and finetuned GPT2 model that generates playable Super Mario Bros levels from text prompts. It accompanies the NeurIPS 2023 paper on open-ended Text2Level generation and includes tools for sampling, rendering, playing, and validating levels.

## Use cases
- generate super mario levels from a text prompt
- procedurally generate game levels with a language model
- research controllable level generation with GPT2
- play or validate AI-generated mario levels with an A* agent
- fine-tune language models on game level data

## When to choose
- you want text-guided procedural level generation for 2D platformers
- you need a reproducible research baseline for open-ended game content generation
- you want a pip-installable library with pretrained Mario level models

## When to avoid
- you need level generation for games other than Super Mario Bros-style platformers
- you need production-grade, polished level output without manual tuning
- you need a general-purpose text-to-game generator

## Facets
- artifact type: library
- maturity: active
- function: machine-learning, llm-inference, data-generation, game-engine
- domain: machine-learning, large-language-models
- platform: python, cross-platform
- tags: procedural-content-generation, gpt2, level-generation, super-mario, research-code, neurips-2023, game-development, gpu

## Member repositories
- shyamsn97/mario-gpt (main) score 21

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:03:46.892078+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T06:33:34.233004+00:00, confidence not recorded.
  - readme: https://github.com/shyamsn97/mario-gpt (fetched 2026-08-28T04:03:46.892078+00:00, sha 435b7ae1bcc8)
  - homepage: https://huggingface.co/shyamsn97/Mario-GPT2-700-context-length (fetched 2026-08-29T12:38:27.126304+00:00, sha 75c27c6dec09)
  - site_page: https://huggingface.co/docs (fetched 2026-08-29T12:38:27.128757+00:00, sha bdec26667b98)
  - site_page: https://huggingface.co/docs/hub/model-cards (fetched 2026-08-29T12:38:27.132939+00:00, sha 60ded09a56b0)
  - site_page: https://huggingface.co/docs/inference-providers (fetched 2026-08-29T12:38:27.134944+00:00, sha 8a5d0f819473)
  - registry_pypi: https://pypi.org/pypi/mario-gpt/json (fetched 2026-08-29T12:38:27.139223+00:00, sha c44fc5a17c95)
  - site_page: https://huggingface.co/pricing (fetched 2026-08-29T12:38:27.130809+00:00, sha de6b7a178be5)
  - site_page: https://huggingface.co/huggingface (fetched 2026-08-29T12:38:27.137104+00:00, sha 59ec3a169a3f)
- Data as of 2026-08-30T08:39:29.467469+00:00.