Model card for hiera_huge_224.mae

A Hiera image feature model. Pretrained on ImageNet-1k with Self-Supervised Masked Autoencoder (MAE) method by paper authors.

Model Details

• Model Type: Image classification / feature backbone
• Model Stats:
  • Params (M): 670.7
  • GMACs: 124.9
  • Activations (M): 151.0
  • Image size: 224 x 224
• Dataset: ImageNet-1k
• Papers:
  • Hiera: A Hierarchical Vision Transformer without the Bells-and-Whistles: https://arxiv.org/abs/2306.00989
  • Masked Autoencoders Are Scalable Vision Learners: https://arxiv.org/abs/2111.06377
• Original: https://github.com/facebookresearch/hiera

Model Usage

Image Classification

from urllib.request import urlopen
from PIL import Image
import timm

img = Image.open(urlopen(
    'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png'
))

model = timm.create_model('hiera_huge_224.mae', pretrained=True)
model = model.eval()

# get model specific transforms (normalization, resize)
data_config = timm.data.resolve_model_data_config(model)
transforms = timm.data.create_transform(**data_config, is_training=False)

output = model(transforms(img).unsqueeze(0))  # unsqueeze single image into batch of 1

top5_probabilities, top5_class_indices = torch.topk(output.softmax(dim=1) * 100, k=5)

Feature Map Extraction

from urllib.request import urlopen
from PIL import Image
import timm

img = Image.open(urlopen(
    'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png'
))

model = timm.create_model(
    'hiera_huge_224.mae',
    pretrained=True,
    features_only=True,
)
model = model.eval()

# get model specific transforms (normalization, resize)
data_config = timm.data.resolve_model_data_config(model)
transforms = timm.data.create_transform(**data_config, is_training=False)

output = model(transforms(img).unsqueeze(0))  # unsqueeze single image into batch of 1

for o in output:
    # print shape of each feature map in output
    # e.g.:
    #  torch.Size([1, 256, 56, 56])
    #  torch.Size([1, 512, 28, 28])
    #  torch.Size([1, 1024, 14, 14])
    #  torch.Size([1, 2048, 7, 7])

    print(o.shape)

Image Embeddings

from urllib.request import urlopen
from PIL import Image
import timm

img = Image.open(urlopen(
    'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png'
))

model = timm.create_model(
    'hiera_huge_224.mae',
    pretrained=True,
    num_classes=0,  # remove classifier nn.Linear
)
model = model.eval()

# get model specific transforms (normalization, resize)
data_config = timm.data.resolve_model_data_config(model)
transforms = timm.data.create_transform(**data_config, is_training=False)

output = model(transforms(img).unsqueeze(0))  # output is (batch_size, num_features) shaped tensor

# or equivalently (without needing to set num_classes=0)

output = model.forward_features(transforms(img).unsqueeze(0))
# output is unpooled, a (1, 49, 2048) shaped tensor

output = model.forward_head(output, pre_logits=True)
# output is a (1, num_features) shaped tensor

Model Comparison

By Top-1

model	top1	top1_err	top5	top5_err	param_count
hiera_huge_224.mae_in1k_ft_in1k	86.834	13.166	98.01	1.99	672.78
hiera_large_224.mae_in1k_ft_in1k	86.042	13.958	97.648	2.352	213.74
hiera_base_plus_224.mae_in1k_ft_in1k	85.134	14.866	97.158	2.842	69.9
hiera_base_224.mae_in1k_ft_in1k	84.49	15.51	97.032	2.968	51.52
hiera_small_224.mae_in1k_ft_in1k	83.884	16.116	96.684	3.316	35.01
hiera_tiny_224.mae_in1k_ft_in1k	82.786	17.214	96.204	3.796	27.91

Citation

@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}
}

@Article{MaskedAutoencoders2021,
  author  = {Kaiming He and Xinlei Chen and Saining Xie and Yanghao Li and Piotr Doll{'a}r and Ross Girshick},
  journal = {arXiv:2111.06377},
  title   = {Masked Autoencoders Are Scalable Vision Learners},
  year    = {2021},
}