metadata

license: mit
tags:
  - vision
  - image-classification
datasets:
  - imagenet-1k
widget:
  - src: >-
      https://huggingface.co/datasets/mishig/sample_images/resolve/main/tiger.jpg
    example_title: Tiger
  - src: >-
      https://huggingface.co/datasets/mishig/sample_images/resolve/main/teapot.jpg
    example_title: Teapot
  - src: >-
      https://huggingface.co/datasets/mishig/sample_images/resolve/main/palace.jpg
    example_title: Palace

NAT (tiny variant)

NAT-Tiny trained on ImageNet-1K at 224x224 resolution. It was introduced in the paper Neighborhood Attention Transformer by Hassani et al. and first released in this repository.

Model description

NAT is a hierarchical vision transformer based on Neighborhood Attention (NA). Neighborhood Attention is a restricted self attention pattern in which each token's receptive field is limited to its nearest neighboring pixels. NA is a sliding-window attention patterns, and as a result is highly flexible and maintains translational equivariance.

NA is implemented in PyTorch implementations through its extension, NATTEN.

Source

Intended uses & limitations

You can use the raw model for image classification. See the model hub to look for fine-tuned versions on a task that interests you.

Example

Here is how to use this model to classify an image from the COCO 2017 dataset into one of the 1,000 ImageNet classes:

from transformers import AutoImageProcessor, NatForImageClassification
from PIL import Image
import requests

url = "http://images.cocodataset.org/val2017/000000039769.jpg"
image = Image.open(requests.get(url, stream=True).raw)

feature_extractor = AutoImageProcessor.from_pretrained("shi-labs/nat-tiny-in1k-224")
model = NatForImageClassification.from_pretrained("shi-labs/nat-tiny-in1k-224")

inputs = feature_extractor(images=image, return_tensors="pt")
outputs = model(**inputs)
logits = outputs.logits
# model predicts one of the 1000 ImageNet classes
predicted_class_idx = logits.argmax(-1).item()
print("Predicted class:", model.config.id2label[predicted_class_idx])

For more examples, please refer to the documentation.

Requirements

Other than transformers, this model requires the NATTEN package.

If you're on Linux, you can refer to shi-labs.com/natten for instructions on installing with pre-compiled binaries (just select your torch build to get the correct wheel URL).

You can alternatively use pip install natten to compile on your device, which may take up to a few minutes. Mac users only have the latter option (no pre-compiled binaries).

Refer to NATTEN's GitHub for more information.

BibTeX entry and citation info

@article{hassani2022neighborhood,
    title        = {Neighborhood Attention Transformer},
    author       = {Ali Hassani and Steven Walton and Jiachen Li and Shen Li and Humphrey Shi},
    year         = 2022,
    url          = {https://arxiv.org/abs/2204.07143},
    eprint       = {2204.07143},
    archiveprefix = {arXiv},
    primaryclass = {cs.CV}
}