🚀 Hiera Model (Tiny, fine-tuned on IN1K)
Hiera is a hierarchical vision transformer that offers high speed, strong performance, and above all, simplicity. It was introduced in the paper Hiera: A Hierarchical Vision Transformer without the Bells-and-Whistles and outperforms the state-of-the-art in a wide range of image and video tasks while being much faster.
Property |
Details |
Datasets |
imagenet-1k |
Language |
en |
Library Name |
transformers |
License |
cc-by-nc-4.0 |
🚀 Quick Start
How does it work?

Vision transformers like ViT maintain the same spatial resolution and number of features throughout the entire network. However, this approach is inefficient as the early layers don't require as many features, and the later layers don't need high spatial resolution. Prior hierarchical models such as ResNet addressed this by using fewer features at the start and less spatial resolution at the end.
Several domain-specific vision transformers with hierarchical designs have been introduced, such as Swin or MViT. In the pursuit of state-of-the-art results through fully supervised training on ImageNet-1K, these models have become increasingly complex as they add specialized modules to compensate for the spatial biases lacking in ViTs. Although these changes result in effective models with appealing FLOP counts, the added complexity actually makes these models slower overall.
We demonstrate that much of this complexity is unnecessary. Instead of manually adding spatial bases through architectural changes, we choose to teach the model these biases. By training with MAE, we can simplify or remove all of these bulky modules in existing transformers and increase accuracy in the process. The outcome is Hiera, an extremely efficient and simple architecture that outperforms the state-of-the-art in several image and video recognition tasks.
✨ Features
Intended uses & limitations
Hiera can be used for image classification, feature extraction, or masked image modeling. This specific checkpoint is intended for Feature Extraction.
💻 Usage Examples
Basic Usage
from transformers import AutoImageProcessor, HieraModel
import torch
from PIL import Image
import requests
url = "http://images.cocodataset.org/val2017/000000039769.jpg"
image = Image.open(requests.get(url, stream=True).raw)
image_processor = AutoImageProcessor.from_pretrained("facebook/hiera-large-224-hf")
model = HieraModel.from_pretrained("facebook/hiera-large-224-hf")
inputs = image_processor(images=image, return_tensors="pt")
outputs = model(**inputs)
Advanced Usage
You can also extract feature maps from different stages of the model using HieraBackbone
and setting out_features
when loading the model. This is how you would extract feature maps from every stage:
from transformers import AutoImageProcessor, HieraBackbone
import torch
from PIL import Image
import requests
url = "http://images.cocodataset.org/val2017/000000039769.jpg"
image = Image.open(requests.get(url, stream=True).raw)
image_processor = AutoImageProcessor.from_pretrained("facebook/hiera-large-224-hf")
model = HieraBackbone.from_pretrained("facebook/hiera-large-224-hf", out_features=['stage1', 'stage2', 'stage3', 'stage4'])
inputs = image_processor(images=image, return_tensors="pt")
outputs = model(**inputs)
feature_maps = outputs.feature_maps
📄 License
BibTeX entry and citation info
If you use Hiera or this code in your work, please cite:
@article{ryali2023hiera,
title={Hiera: A Hierarchical Vision Transformer without the Bells-and-Whistles},
author={Ryali, Chaitanya and Hu, Yuan-Ting and Bolya, Daniel and Wei, Chen and Fan, Haoqi and Huang, Po-Yao and Aggarwal, Vaibhav and Chowdhury, Arkabandhu and Poursaeed, Omid and Hoffman, Judy and Malik, Jitendra and Li, Yanghao and Feichtenhofer, Christoph},
journal={ICML},
year={2023}
}