#318 · Primary category： Computer Vision
EfficientFormer
EfficientFormerV2 [ICCV 2023] & EfficientFormer [NeurIPs 2022]
Project last updated：08/13/23
GitHub Stars
1.1K
Forks
95
Contributors
2
License
Other
Why we included this project
EfficientFormer is Snap Research's official release for a family of vision transformers built around mobile deployment. The V2 revision, presented at ICCV 2023, targets MobileNet-scale latency and parameter counts; on ImageNet-1K it lands roughly 4% higher top-1 accuracy than MobileNetV2 at similar speed and size. The repo ships pretrained checkpoints and measured iPhone 12 latency figures, so the accuracy-versus-speed trade-off is concrete rather than a guess. It is also worth reading if you care about how the compact designs were found: the authors describe a fine-grained joint search that optimizes latency and parameter count together, which is more instructive than handing over a finished model. Teams weighing transformer backbones for on-device inference will find the weights and deployment guidance in one place.
Articles for this project
No articles for this project yet.
To suggest a topic or contribute an article, contact us.
Related projects in this category
opencv
Open Source Computer Vision Library
RuView
π RuView turns commodity WiFi signals into real-time spatial intelligence, vital sign monitoring, and presence detection — all without a single pixel of video.
PaddleOCR
Turn any PDF or image document into structured data for your AI. A powerful, lightweight OCR toolkit that bridges the gap between images/PDFs and LLMs. Supports 100+ languages.
MinerU
Transforms complex documents like PDFs and Office docs into LLM-ready markdown/JSON for your Agentic workflows.
tesseract
Tesseract Open Source OCR Engine (main repository)