Model card for mambaout_base_plus_rw.sw_e150_r384_in12k_ft_in1k

A MambaOut image classification model with timm specific architecture customizations. Pretrained on ImageNet-12k and fine-tuned on ImageNet-1k by Ross Wightman using Swin / ConvNeXt based recipe.

Model Details

• Model Type: Image classification / feature backbone
• Model Stats:
  • Params (M): 101.7
  • GMACs: 56.4
  • Activations (M): 132.7
  • Image size: 384 x 384
• Pretrain Dataset: ImageNet-12k
• Dataset: ImageNet-1k
• Papers:
  • PyTorch Image Models: https://github.com/huggingface/pytorch-image-models
  • MambaOut: Do We Really Need Mamba for Vision?: https://arxiv.org/abs/2405.07992
• Original: https://github.com/yuweihao/MambaOut

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('mambaout_base_plus_rw.sw_e150_r384_in12k_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

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(
    'mambaout_base_plus_rw.sw_e150_r384_in12k_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, 96, 128])
    #  torch.Size([1, 48, 48, 256])
    #  torch.Size([1, 24, 24, 512])
    #  torch.Size([1, 12, 12, 768])

    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(
    'mambaout_base_plus_rw.sw_e150_r384_in12k_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, 12, 12, 768) shaped tensor

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

Model Comparison

By Top-1

Citation

@misc{rw2019timm,
  author = {Ross Wightman},
  title = {PyTorch Image Models},
  year = {2019},
  publisher = {GitHub},
  journal = {GitHub repository},
  doi = {10.5281/zenodo.4414861},
  howpublished = {\url{https://github.com/huggingface/pytorch-image-models}}
}

@article{yu2024mambaout,
  title={MambaOut: Do We Really Need Mamba for Vision?},
  author={Yu, Weihao and Wang, Xinchao},
  journal={arXiv preprint arXiv:2405.07992},
  year={2024}
}