# snap-research/EfficientFormer

EfficientFormerV2 [ICCV 2023] & EfficientFormer [NeurIPs 2022]

Repository: https://github.com/snap-research/EfficientFormer
Canonical: https://ross.abutalabs.com/products/efficientformer
Homepage: https://arxiv.org/abs/2212.08059
Language: Python
License: NOASSERTION
License Family: other
Topics: deep-learning, detection, efficient-inference, efficient-neural-networks, pytorch, semantic-segmentation, transformer, imagenet, transformers, mobile-devices
Last push: 2023-08-13T13:28:39+00:00

## Health v2 (maintenance only)
Score: 32/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 0, release rhythm 35, longevity 100
- inputs: {"age_days": 1553, "days_push": 1116, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases, no_license
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1116, forks 95 (observed 2026-08-28T04:03:38.775164+00:00)

## What it is
A PyTorch implementation of EfficientFormer and EfficientFormerV2, efficient vision transformer model families designed to run at MobileNet-level speed and size on mobile devices. It includes pretrained ImageNet-1K checkpoints, downstream detection/segmentation variants, and profiling tools.

## Use cases
- run vision transformers as fast as mobilenet on iphone
- download pretrained efficientformer checkpoints for imagenet classification
- use efficient vision transformer backbone for object detection
- use efficientformer for semantic segmentation
- benchmark transformer latency on mobile devices
- find efficient transformer architectures for resource-constrained hardware

## When to choose
- you need low-latency image classification on mobile or edge devices
- you want a vision transformer backbone with MobileNet-level size and speed
- you need pretrained checkpoints for detection or segmentation with efficient backbones

## When to avoid
- you need state-of-the-art accuracy on high-end GPUs regardless of latency
- you need a maintained library with frequent updates or broad ecosystem support
- you need non-PyTorch frameworks like TensorFlow or JAX

## Facets
- artifact type: library
- maturity: maintenance
- function: machine-learning, deep-learning, image-processing, computer-vision
- domain: deep-learning, computer-vision, image-processing, mobile-development
- platform: python, cross-platform
- tags: vision-transformer, efficient-inference, imagenet, pytorch, model-zoo, neural-architecture-search, mobile-deployment, semantic-segmentation, object-detection, mobile

## Member repositories
- snap-research/EfficientFormer (main) score 32

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:03:38.775164+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 2026-08-30T06:41:59.810947+00:00, confidence not recorded.
  - readme: https://github.com/snap-research/EfficientFormer (fetched 2026-08-28T04:03:38.775164+00:00, sha 7d0668b7003c)
  - homepage: https://arxiv.org/abs/2212.08059 (fetched 2026-08-29T12:46:10.770495+00:00, sha d6ab4d76f779)
  - site_page: https://info.arxiv.org/about/donate.html (fetched 2026-08-29T12:46:10.779903+00:00, sha cca9c3a11c56)
  - site_page: https://info.arxiv.org/about/ourmembers.html (fetched 2026-08-29T12:46:10.783365+00:00, sha 47cbc55ff1de)
  - site_page: https://info.arxiv.org/about (fetched 2026-08-29T12:46:10.785549+00:00, sha a1f16f915a9a)
  - site_page: https://info.arxiv.org/labs/index.html (fetched 2026-08-29T12:46:10.781643+00:00, sha b14a8d05a0ec)
- Data as of 2026-08-30T08:39:29.467469+00:00.