snap-research/EfficientFormer
A vision transformer optimized to match MobileNet speed and size for efficient mobile deployment.
Not currently ranked — collecting fresh signals.
star history
EfficientFormer is a family of efficient vision transformers designed for mobile and edge deployment. The work investigates whether transformer models can achieve latency and parameter counts comparable to lightweight CNNs like MobileNet. The models include attention mechanisms and hybrid architectures optimized for resource-constrained hardware, with inference实测 on iPhone 12 using CoreMLTools.
Frequently asked
- What is snap-research/EfficientFormer?
- A vision transformer optimized to match MobileNet speed and size for efficient mobile deployment.
- Is EfficientFormer open source?
- Yes — snap-research/EfficientFormer is an open-source project tracked on heatdrop.
- What language is EfficientFormer written in?
- snap-research/EfficientFormer is primarily written in Python.
- How popular is EfficientFormer?
- snap-research/EfficientFormer has 1.1k stars on GitHub.
- Where can I find EfficientFormer?
- snap-research/EfficientFormer is on GitHub at https://github.com/snap-research/EfficientFormer.