# openai/pixel-cnn

Code for the paper "PixelCNN++: A PixelCNN Implementation with Discretized Logistic Mixture Likelihood and Other Modifications"

Repository: https://github.com/openai/pixel-cnn
Canonical: https://ross.abutalabs.com/products/pixel-cnn
Homepage: https://arxiv.org/abs/1701.05517
Language: Python
License: NOASSERTION
License Family: other
Topics: paper
Archived: true
Last push: 2020-02-03T06:39:20+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": 3708, "days_push": 2403, "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 1960, forks 431 (observed 2026-08-28T04:05:59.364191+00:00)

## What it is
A Python/TensorFlow implementation of PixelCNN++, an autoregressive generative model for images with discretized logistic mixture likelihood, from the OpenAI paper of the same name. It supports multi-GPU training on CIFAR-10 and Small ImageNet and includes a pretrained checkpoint achieving 2.92 bits per dimension.

## Use cases
- train a PixelCNN generative model on CIFAR-10
- generate images from an autoregressive pixel model
- reproduce PixelCNN++ paper results
- compute tractable image likelihoods in bits per dimension
- train a class-conditional image generation model
- adapt a generative image model to my own dataset

## When to choose
- you need the exact PixelCNN++ model from the paper for research reproduction
- you want a TensorFlow autoregressive image model with tractable likelihood
- you have multi-GPU hardware and want to train on CIFAR-10 or Small ImageNet

## When to avoid
- you need maintained code or recent TensorFlow compatibility
- you want modern generative models like diffusion or transformers
- you only have a single GPU or no GPU
- you need a production-ready image generation service

## Facets
- artifact type: library
- maturity: abandoned
- function: machine-learning, deep-learning, image-processing
- domain: machine-learning, deep-learning, computer-vision
- platform: python
- tags: generative-models, pixelcnn, tensorflow, research-code, image-generation, archived, gpu, linux

## Member repositories
- openai/pixel-cnn (main) score 10

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:05:59.364191+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 2026-08-30T03:05:48.225073+00:00, confidence not recorded.
  - readme: https://github.com/openai/pixel-cnn (fetched 2026-08-28T04:05:59.364191+00:00, sha 944cb287acbe)
  - homepage: https://arxiv.org/abs/1701.05517 (fetched 2026-08-29T10:45:50.642697+00:00, sha 43c38985a0c8)
  - site_page: https://info.arxiv.org/about/donate.html (fetched 2026-08-29T10:45:50.652422+00:00, sha cca9c3a11c56)
  - site_page: https://info.arxiv.org/about/ourmembers.html (fetched 2026-08-29T10:45:50.656226+00:00, sha 47cbc55ff1de)
  - site_page: https://info.arxiv.org/about (fetched 2026-08-29T10:45:50.658510+00:00, sha a1f16f915a9a)
  - site_page: https://info.arxiv.org/labs/index.html (fetched 2026-08-29T10:45:50.654453+00:00, sha b14a8d05a0ec)
- Data as of 2026-08-30T08:39:29.467469+00:00.