timm
/

Image Classification
timm
PyTorch
Safetensors
rwightman's picture
rwightman HF staff
Update model config and README
2f28011 verified
---
tags:
- image-classification
- timm
library_name: timm
license: cc-by-nc-4.0
datasets:
- imagenet-1k
---
# Model card for hiera_small_224.mae_in1k_ft_in1k
A Hiera image classification model. Pretrained on ImageNet-1k with Self-Supervised Masked Autoencoder (MAE) method and fine-tuned on ImageNet-1k.
## Model Details
- **Model Type:** Image classification / feature backbone
- **Model Stats:**
- Params (M): 35.0
- GMACs: 6.0
- Activations (M): 17.3
- 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
```python
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_small_224.mae_in1k_ft_in1k', 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
```python
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_small_224.mae_in1k_ft_in1k',
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, 96, 56, 56])
# torch.Size([1, 192, 28, 28])
# torch.Size([1, 384, 14, 14])
# torch.Size([1, 768, 7, 7])
print(o.shape)
```
### Image Embeddings
```python
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_small_224.mae_in1k_ft_in1k',
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, 768) 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 |top5 |param_count|
|---------------------------------|------|------|-----------|
|hiera_huge_224.mae_in1k_ft_in1k |86.834|98.01 |672.78 |
|hiera_large_224.mae_in1k_ft_in1k |86.042|97.648|213.74 |
|hiera_base_plus_224.mae_in1k_ft_in1k|85.134|97.158|69.9 |
|hiera_small_abswin_256.sbb2_e200_in12k_ft_in1k |84.912|97.260|35.01 |
|hiera_small_abswin_256.sbb2_pd_e200_in12k_ft_in1k |84.560|97.106|35.01 |
|hiera_base_224.mae_in1k_ft_in1k |84.49 |97.032|51.52 |
|hiera_small_224.mae_in1k_ft_in1k |83.884|96.684|35.01 |
|hiera_tiny_224.mae_in1k_ft_in1k |82.786|96.204|27.91 |
## Citation
```bibtex
@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}
}
```
```bibtex
@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},
}
```