模型卡片：nfnet_l0.ra2_in1k

一种NFNet-Lite（轻量级NFNet）图像分类模型。由Ross Wightman在timm中训练。

无标准化网络（Normalization Free Networks）是（预激活的）ResNet类似的模型，没有任何标准化层。它们使用了缩放权重标准化和根据信号传播分析在残差路径和非线性函数处放置特定的标量增益，而不是批量归一化或其他替代方法。

轻量级NFNet是timm特定的变种，将SE和瓶颈比从0.5 -> 0.25（减少宽度），并使用较小的组大小，同时保持相同的深度。使用SiLU激活函数而不是GELU。

模型细节

• 模型类型：图像分类/特征骨干
• 模型统计：
  • 参数（M）：35.1
  • GMACs：4.4
  • 激活值（M）：10.5
  • 图像大小：训练 = 224 x 224，测试 = 288 x 288
• 论文：
  • 无标准化的高性能大规模图像识别： https://arxiv.org/abs/2102.06171
  • 表征信号传播以消除非标准化ResNet中的性能差距： https://arxiv.org/abs/2101.08692
• 原始模型： https://github.com/huggingface/pytorch-image-models
• 数据集：ImageNet-1k

模型用法

图像分类

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('nfnet_l0.ra2_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)

特征图提取

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(
    'nfnet_l0.ra2_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, 64, 112, 112])
    #  torch.Size([1, 256, 56, 56])
    #  torch.Size([1, 512, 28, 28])
    #  torch.Size([1, 1536, 14, 14])
    #  torch.Size([1, 2304, 7, 7])

    print(o.shape)

图像嵌入

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(
    'nfnet_l0.ra2_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, 2304, 7, 7) shaped tensor

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

模型比较

在timm的 model results 中探索此模型的数据集和运行时指标。

引用

@article{brock2021high,
  author={Andrew Brock and Soham De and Samuel L. Smith and Karen Simonyan},
  title={High-Performance Large-Scale Image Recognition Without Normalization},
  journal={arXiv preprint arXiv:2102.06171},
  year={2021}
}

@inproceedings{brock2021characterizing,
  author={Andrew Brock and Soham De and Samuel L. Smith},
  title={Characterizing signal propagation to close the performance gap in
  unnormalized ResNets},
  booktitle={9th International Conference on Learning Representations, {ICLR}},
  year={2021}
}

@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}}
}