NVlabs/FasterViT
A vision transformer architecture with hierarchical attention mechanisms for efficient image understanding tasks.
Not currently ranked — collecting fresh signals.
star history
FasterViT is a PyTorch implementation of a hierarchical vision transformer that introduces Hierarchical Attention (HAT) to capture both short and long-range spatial information via cross-window carrier tokens. It serves as a strong backbone for various computer vision tasks including image classification on ImageNet, object detection on COCO, and semantic segmentation on ADE20K. The model achieves competitive accuracy-throughput tradeoffs on standard benchmarks.
Frequently asked
- What is NVlabs/FasterViT?
- A vision transformer architecture with hierarchical attention mechanisms for efficient image understanding tasks.
- Is FasterViT open source?
- Yes — NVlabs/FasterViT is an open-source project tracked on heatdrop.
- What language is FasterViT written in?
- NVlabs/FasterViT is primarily written in Python.
- How popular is FasterViT?
- NVlabs/FasterViT has 921 stars on GitHub.
- Where can I find FasterViT?
- NVlabs/FasterViT is on GitHub at https://github.com/NVlabs/FasterViT.