# FoundationVision/Infinity

[CVPR 2025 Oral]Infinity ∞ : Scaling Bitwise AutoRegressive Modeling for High-Resolution Image Synthesis

Repository: https://github.com/FoundationVision/Infinity
Canonical: https://ross.abutalabs.com/products/foundationvision-infinity
Language: Python
License: MIT
License Family: permissive
Topics: auto-regressive-model, autoregressive-models, generative-model, gpt, gpt-2, image-generation, text-to-image, text-to-image-generation, transformers
Last push: 2026-04-16T03:02:02+00:00

## Health v2 (maintenance only)
Score: 56/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 77, release rhythm 35, longevity 45
- inputs: {"age_days": 642, "days_push": 139, "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 1587, forks 93 (observed 2026-08-28T04:05:07.872394+00:00)

## What it is
Infinity is a bitwise autoregressive text-to-image generation model (CVPR 2025 Oral) with released training and inference code, checkpoints up to 8B parameters, and an image tokenizer (BitVAE). It enables high-resolution image synthesis via visual autoregressive modeling.

## Use cases
- generate high-resolution images from text prompts
- train a text-to-image autoregressive model
- train an image tokenizer like BitVAE
- run interactive text-to-image inference in a notebook
- research visual autoregressive modeling as an alternative to diffusion

## When to choose
- you want autoregressive (non-diffusion) text-to-image generation
- you need open weights and training code for high-resolution image synthesis
- you are researching visual autoregressive or bitwise token modeling

## When to avoid
- you need a lightweight CPU-only image generator
- you want a polished end-user product rather than research code
- you lack access to high-end GPUs for inference or training

## Facets
- artifact type: library
- maturity: active
- function: machine-learning, deep-learning, image-processing, transformers, llm-training
- domain: deep-learning, image-processing, artificial-intelligence, computer-vision
- platform: python
- tags: text-to-image, autoregressive-modeling, image-synthesis, var, bitwise-tokenizer, generative-model, cvpr-2025, gpu, linux

## Member repositories
- FoundationVision/Infinity (main) score 56

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:05:07.872394+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 2026-08-30T03:55:21.626757+00:00, confidence not recorded.
  - readme: https://github.com/FoundationVision/Infinity (fetched 2026-08-28T04:05:07.872394+00:00, sha c1e2c8ffcc29)
- Data as of 2026-08-30T08:39:29.467469+00:00.