# LTH14/JiT

PyTorch implementation of JiT https://arxiv.org/abs/2511.13720

Repository: https://github.com/LTH14/JiT
Canonical: https://ross.abutalabs.com/products/jit
Language: Python
License: MIT
License Family: permissive
Last push: 2025-12-08T07:11:31+00:00

## Health v2 (maintenance only)
Score: 42/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 56, release rhythm 35, longevity 21
- inputs: {"age_days": 296, "days_push": 268, "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 2507, forks 169 (observed 2026-08-28T04:06:57.170574+00:00)

## What it is
A PyTorch/GPU re-implementation of JiT (Just image Transformer), a minimalist pixel-space diffusion model for high-resolution image generation from the paper 'Back to Basics: Let Denoising Generative Models Denoise'. It includes training and evaluation scripts for ImageNet at 256x256 and 512x512 resolutions.

## Use cases
- train a pixel-space diffusion transformer on ImageNet
- generate high-resolution images with a diffusion model
- reproduce the JiT paper results in PyTorch
- experiment with denoising generative models
- train image generation models on multiple GPUs

## When to choose
- you want a simple, self-contained PyTorch implementation of pixel-space diffusion
- you need to reproduce or build on the JiT paper
- you have multi-GPU resources for training image generation models

## When to avoid
- you need a production-ready image generation API
- you lack the GPU resources to train large diffusion models
- you want latent diffusion rather than pixel-space diffusion

## Facets
- artifact type: library
- maturity: active
- function: deep-learning, machine-learning, image-processing
- domain: deep-learning, machine-learning, image-processing
- platform: python
- tags: diffusion-models, image-generation, pytorch, research-code, transformer, pixel-space-diffusion, gpu

## Member repositories
- LTH14/JiT (main) score 42

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:06:57.170574+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 2026-08-30T02:26:33.076789+00:00, confidence not recorded.
  - readme: https://github.com/LTH14/JiT (fetched 2026-08-28T04:06:57.170574+00:00, sha e6cd5141316f)
- Data as of 2026-08-30T08:39:29.467469+00:00.