# jeonsworld/ViT-pytorch

Pytorch reimplementation of the Vision Transformer (An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale)

Repository: https://github.com/jeonsworld/ViT-pytorch
Canonical: https://ross.abutalabs.com/products/jeonsworld-vit-pytorch
Language: Jupyter Notebook
License: MIT
License Family: permissive
Last push: 2022-06-07T07:14:33+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": 2129, "days_push": 1548, "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 2162, forks 397 (observed 2026-08-28T04:06:21.080121+00:00)

## What it is
A PyTorch reimplementation of Google's Vision Transformer (ViT) for image classification, based on the paper 'An Image is Worth 16x16 Words'. It includes scripts to convert official checkpoints, fine-tune on datasets like CIFAR-10/100, and train with mixed precision.

## Use cases
- fine-tune a vision transformer on cifar-10
- load pretrained ViT weights in pytorch
- train an image classifier with transformers
- convert google vit npz checkpoints to pytorch
- experiment with patch-based transformer models for images

## When to choose
- you want a faithful PyTorch port of the original ViT with official pretrained weights
- you need a simple training script for image classification with transformers

## When to avoid
- you need production-grade, actively maintained vision model tooling
- you want a broad model zoo beyond ViT variants

## Facets
- artifact type: library
- maturity: maintenance
- function: machine-learning, deep-learning, image-processing
- domain: deep-learning, computer-vision, image-processing, machine-learning
- platform: python
- tags: vision-transformer, pytorch, image-classification, pretrained-models, transformer, gpu

## Member repositories
- jeonsworld/ViT-pytorch (main) score 32

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:06:21.080121+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 2026-08-30T02:50:14.308583+00:00, confidence not recorded.
  - readme: https://github.com/jeonsworld/ViT-pytorch (fetched 2026-08-28T04:06:21.080121+00:00, sha 431363efdca7)
- Data as of 2026-08-30T08:39:29.467469+00:00.