# PixArt-alpha/PixArt-sigma

PixArt-Σ: Weak-to-Strong Training of Diffusion Transformer for 4K Text-to-Image Generation

Repository: https://github.com/PixArt-alpha/PixArt-sigma
Canonical: https://ross.abutalabs.com/products/pixart-sigma
Homepage: https://pixart-alpha.github.io/PixArt-sigma-project/
Language: Python
License: Apache-2.0
License Family: permissive
Last push: 2024-10-31T01:20:19+00:00

## Health v2 (maintenance only)
Score: 25/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 0, release rhythm 35, longevity 65
- inputs: {"age_days": 916, "days_push": 672, "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 1939, forks 99 (observed 2026-08-28T04:05:57.263017+00:00)

## What it is
PixArt-Σ is a PyTorch implementation of a diffusion transformer model for high-resolution (up to 4K) text-to-image generation, trained with a weak-to-strong training approach. It provides pre-trained checkpoints, inference/sampling code, LoRA training, and HuggingFace diffusers integration.

## Use cases
- generate 4k images from text prompts
- run text-to-image inference with a diffusion transformer
- train LoRA adapters for custom image styles
- generate high-resolution artwork from descriptions
- integrate a text-to-image model into diffusers pipelines
- distill a diffusion model into a one-step generator

## When to avoid
- you need image editing or inpainting rather than text-to-image generation
- you lack a GPU or need lightweight on-device generation
- you need a production-grade managed image generation service
- you need video or audio generation

## Facets
- artifact type: library
- maturity: active
- function: machine-learning, deep-learning, image-processing, llm-training
- domain: artificial-intelligence, deep-learning, image-processing, machine-learning
- platform: python
- tags: text-to-image, diffusion-transformer, diffusion-models, 4k-image-generation, pytorch, stable-diffusion-alternative, lora, research-code, gpu, linux

## Member repositories
- PixArt-alpha/PixArt-sigma (main) score 25

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:05:57.263017+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 2026-08-30T03:08:10.657588+00:00, confidence not recorded.
  - readme: https://github.com/PixArt-alpha/PixArt-sigma (fetched 2026-08-28T04:05:57.263017+00:00, sha 14dc3693d4ca)
  - homepage: https://pixart-alpha.github.io/PixArt-sigma-project/ (fetched 2026-08-29T10:47:25.128001+00:00, sha 6509e5c22807)
- Data as of 2026-08-30T08:39:29.467469+00:00.