# minimaxir/gpt-2-simple

Python package to easily retrain OpenAI's GPT-2 text-generating model on new texts

Repository: https://github.com/minimaxir/gpt-2-simple
Canonical: https://ross.abutalabs.com/products/gpt-2-simple
Language: Python
License: NOASSERTION
License Family: other
Topics: text-generation, tensorflow, openai, textgenrnn
Last push: 2022-12-14T11:50:45+00:00

## Health v2 (maintenance only)
Score: 23/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 0, release rhythm 8, longevity 100
- inputs: {"age_days": 2699, "days_push": 1358, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_license
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 3400, forks 666 (observed 2026-08-28T04:08:02.796931+00:00)

## What it is
A Python package that simplifies fine-tuning OpenAI's GPT-2 text-generation model (124M/355M) on custom text and generating text from the retrained model, built on TensorFlow. Development has largely been superseded by the same author's aitextgen.

## Use cases
- fine-tune gpt-2 on my own text corpus
- generate text with a retrained gpt-2 model
- train a small language model on a custom dataset
- generate text continuations from a prompt prefix
- run gpt-2 training in a colab notebook

## When to choose
- you specifically need TensorFlow-based GPT-2 fine-tuning
- you want a simple, scriptable API for retraining and generating with GPT-2
- you have existing gpt-2-simple checkpoints to continue using

## When to avoid
- you want modern, efficient LLM fine-tuning - use aitextgen or Hugging Face transformers instead
- you need models larger than GPT-2 medium (355M)
- you want actively developed features or PyTorch support

## Facets
- artifact type: library
- maturity: maintenance
- function: llm-training, llm-inference, machine-learning, deep-learning
- domain: large-language-models, deep-learning, machine-learning
- platform: python, cross-platform
- tags: gpt-2, fine-tuning, text-generation, tensorflow, superseded-by-aitextgen, natural-language-processing, gpu

## Member repositories
- minimaxir/gpt-2-simple (main) score 23

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:08:02.796931+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 2026-08-29T18:38:42.714570+00:00, confidence not recorded.
  - readme: https://github.com/minimaxir/gpt-2-simple (fetched 2026-08-28T04:08:02.796931+00:00, sha 74e8a5971b3e)
  - registry_pypi: https://pypi.org/pypi/gpt-2-simple/json (fetched 2026-08-29T09:32:55.638222+00:00, sha fec1b3b1b82a)
- Data as of 2026-08-30T08:39:29.467469+00:00.