# openai/image-gpt

Repository: https://github.com/openai/image-gpt
Canonical: https://ross.abutalabs.com/products/image-gpt
Language: Python
License: NOASSERTION
License Family: other
Archived: true
Last push: 2022-04-29T00:03:56+00:00

## Health v2 (maintenance only)
Score: 10/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 0, release rhythm 35, longevity 100
- inputs: {"age_days": 2309, "days_push": 1588, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases, archived, no_license
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 2096, forks 387 (observed 2026-08-28T04:06:13.277375+00:00)

## What it is
OpenAI's official code and pre-trained models for iGPT (image GPT), a GPT-2-style transformer adapted to generate and classify images as pixel sequences. The repository is archived and provided as-is with no further updates.

## Use cases
- generate images with a pretrained transformer model
- experiment with autoregressive image modeling
- pretrain transformers on pixels for representation learning
- evaluate iGPT on imagenet and cifar10
- sample from iGPT-S/M/L checkpoints
- study how GPT-2 architecture applies to images

## When to choose
- you need the exact iGPT models from the Generative Pretraining from Pixels paper
- you want a research starting point for autoregressive pixel-based image generation
- you can work with TensorFlow 1.x on Ubuntu with multi-GPU setups

## When to avoid
- you need maintained code or modern framework support
- you want production image generation rather than research experimentation
- you cannot use TensorFlow 1.13 or legacy CUDA environments

## Facets
- artifact type: library
- maturity: abandoned
- function: machine-learning, deep-learning, llm-training, image-processing
- domain: deep-learning, machine-learning, computer-vision, artificial-intelligence
- platform: python
- tags: gpt, transformers, image-generation, tensorflow, pretrained-models, research, openai, archived, linux, gpu

## Member repositories
- openai/image-gpt (main) score 10

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:06:13.277375+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 2026-08-30T02:54:37.806969+00:00, confidence not recorded.
  - readme: https://github.com/openai/image-gpt (fetched 2026-08-28T04:06:13.277375+00:00, sha 477654da6192)
- Data as of 2026-08-30T08:39:29.467469+00:00.