# zai-org/CogView

Text-to-Image generation. The repo for NeurIPS 2021 paper "CogView: Mastering Text-to-Image Generation via Transformers".

Repository: https://github.com/zai-org/CogView
Canonical: https://ross.abutalabs.com/products/cogview
Language: Python
License: Apache-2.0
License Family: permissive
Topics: text-to-image, transformers, pretrained-models, pytorch
Last push: 2023-09-25T04:07:19+00:00

## Health v2 (maintenance only)
Score: 32/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 0, release rhythm 35, longevity 100
- inputs: {"age_days": 1926, "days_push": 1073, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1799, forks 175 (observed 2026-08-28T04:05:37.894475+00:00)

## What it is
CogView is a 4-billion-parameter pretrained transformer model for text-to-image image generation, released alongside a NeurIPS 2021 paper. It provides PyTorch code for inference and training, with a focus on Chinese text input (English supported via translation).

## Use cases
- generate images from Chinese text descriptions
- run a pretrained text-to-image transformer model
- train large deep transformers stably without NaN losses
- experiment with text-to-image generation research
- set up text-to-image generation on Google Colab

## When to choose
- you need a pretrained transformer-based text-to-image model, especially for Chinese prompts
- you want to reproduce or build on the CogView NeurIPS 2021 research
- you need techniques like PB-relax or Sandwich-LN for stable training of large transformers

## When to avoid
- you need state-of-the-art image quality - newer models like CogView2 or modern diffusion models perform better
- you need English-first text-to-image generation without translation
- you lack access to V100/A100 GPUs or similar hardware
- you want an actively developed production tool rather than a research artifact

## Facets
- artifact type: library
- maturity: maintenance
- function: machine-learning, deep-learning, llm-training
- domain: artificial-intelligence, image-processing, deep-learning
- platform: python
- tags: text-to-image, transformers, pretrained-models, image-generation, pytorch, research, natural-language-processing, linux, gpu

## Member repositories
- zai-org/CogView (main) score 32

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:05:37.894475+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 2026-08-30T03:22:18.748685+00:00, confidence not recorded.
  - readme: https://github.com/zai-org/CogView (fetched 2026-08-28T04:05:37.894475+00:00, sha bb7b63022cdf)
- Data as of 2026-08-30T08:39:29.467469+00:00.