# BR-IDL/PaddleViT

:robot: PaddleViT: State-of-the-art Visual Transformer and MLP Models for PaddlePaddle 2.0+

Repository: https://github.com/BR-IDL/PaddleViT
Canonical: https://ross.abutalabs.com/products/paddlevit
Homepage: https://github.com/BR-IDL/PaddleViT
Language: Python
License: Apache-2.0
License Family: permissive
Topics: cv, computer-vision, paddlepaddle, vit, mlp, transformer, encoder-decoder, classification, detection, segmentation, gan, deep-learning, semantic-segmentation, object-detection
Last push: 2022-09-07T08:37:15+00:00

## Health v2 (maintenance only)
Score: 23/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 0, release rhythm 8, longevity 100
- inputs: {"age_days": 1829, "days_push": 1456, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1238, forks 328 (observed 2026-08-28T04:04:05.605630+00:00)

## What it is
PaddleViT is a collection of state-of-the-art Vision Transformer and MLP model implementations for PaddlePaddle 2.1+, covering image classification, object detection, semantic segmentation, GANs, and self-supervised learning. It includes pretrained weights, training/validation scripts, data augmentations, optimizers, and utilities to support research and deployment.

## Use cases
- train vision transformer models for image classification
- fine-tune pretrained ViT models on my own dataset
- run object detection with DETR in PaddlePaddle
- do semantic segmentation with transformer-based models
- reproduce state-of-the-art ViT and MLP papers
- export vision transformer models for production inference
- learn vision transformers with tutorials

## When to choose
- you use PaddlePaddle and want transformer-based vision models
- you need pretrained ViT/MLP weights for fine-tuning
- you want a research-friendly codebase with standalone model modules
- you need multiple CV tasks (classification, detection, segmentation, GAN) in one collection

## When to avoid
- your stack is PyTorch or TensorFlow rather than PaddlePaddle
- you need production-grade serving infrastructure out of the box
- you require very recent model architectures, as the last release was in 2022

## Facets
- artifact type: library
- maturity: maintenance
- function: deep-learning, machine-learning, image-processing, computer-vision
- domain: computer-vision, deep-learning, machine-learning, image-processing
- platform: python, cross-platform
- tags: vision-transformer, vit, paddlepaddle, mlp-mixer, image-classification, object-detection, semantic-segmentation, gan, pretrained-models, model-zoo, gpu

## Member repositories
- BR-IDL/PaddleViT (main) score 23

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:04:05.605630+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 2026-08-30T05:10:01.070637+00:00, confidence not recorded.
  - readme: https://github.com/BR-IDL/PaddleViT (fetched 2026-08-28T04:04:05.605630+00:00, sha e2051021fb35)
  - homepage: https://github.com/BR-IDL/PaddleViT (fetched 2026-08-29T12:21:00.975677+00:00, sha 5ff8ca1ce78f)
- Data as of 2026-08-30T08:39:29.467469+00:00.