mobile-vit-xxs / README.md
merve's picture
merve HF staff
Set `library_name` to `tf-keras`. (#1)
fa60ea7 verified
metadata
library_name: tf-keras
license:
  - cc0-1.0
tags:
  - computer-vision
  - image-classification

Image Classification using MobileViT

This repo contains the model and the notebook to this Keras example on MobileViT.

Full credits to: Sayak Paul

Background Information

MobileViT architecture (Mehta et al.), combines the benefits of Transformers (Vaswani et al.) and convolutions. With Transformers, we can capture long-range dependencies that result in global representations. With convolutions, we can capture spatial relationships that model locality.

Besides combining the properties of Transformers and convolutions, the authors introduce MobileViT as a general-purpose mobile-friendly backbone for different image recognition tasks. Their findings suggest that, performance-wise, MobileViT is better than other models with the same or higher complexity (MobileNetV3, for example), while being efficient on mobile devices.

Training Data

The model is trained on a tf_flowers dataset