inception_resnet_v2.tf_ens_adv_in1k 模型卡片

一个 Inception-ResNet-v2 图像分类模型。由论文作者在 ImageNet-1k 上进行了对抗训练（集成）。由 Ross Wightman 进行了从 Tensorflow 的移植。

模型详情

• 模型类型： 图像分类 / 特征骨干
• 模型统计数据：
  • 参数数量（M）：55.8
  • GMACs（十亿次乘加运算）：13.2
  • 激活函数数量（M）：25.1
  • 图像尺寸：299 x 299
• 论文：
  • https://arxiv.org/abs/1602.07261 : https://arxiv.org/abs/1602.07261
  • 对抗攻击和防御竞赛： https://arxiv.org/abs/1804.00097
• 原始来源： https://github.com/tensorflow/models
• 数据集： ImageNet-1k

模型用途

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('inception_resnet_v2.tf_ens_adv_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(
    'inception_resnet_v2.tf_ens_adv_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, 147, 147])
    #  torch.Size([1, 192, 71, 71])
    #  torch.Size([1, 320, 35, 35])
    #  torch.Size([1, 1088, 17, 17])
    #  torch.Size([1, 1536, 8, 8])

    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(
    'inception_resnet_v2.tf_ens_adv_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, 1536, 8, 8) shaped tensor

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

模型比较

在 timm（代码库）中探索该模型数据集和运行时指标 model results 。

引用

@article{Szegedy2016Inceptionv4IA,
  title={Inception-v4, Inception-ResNet and the Impact of Residual Connections on Learning},
  author={Christian Szegedy and Sergey Ioffe and Vincent Vanhoucke and Alexander A. Alemi},
  journal={ArXiv},
  year={2016},
  volume={abs/1602.07261}
}

@article{Kurakin2018AdversarialAA,
  title={Adversarial Attacks and Defences Competition},
  author={Alexey Kurakin and Ian J. Goodfellow and Samy Bengio and Yinpeng Dong and Fangzhou Liao and Ming Liang and Tianyu Pang and Jun Zhu and Xiaolin Hu and Cihang Xie and Jianyu Wang and Zhishuai Zhang and Zhou Ren and Alan Loddon Yuille and Sangxia Huang and Yao Zhao and Yuzhe Zhao and Zhonglin Han and Junjiajia Long and Yerkebulan Berdibekov and Takuya Akiba and Seiya Tokui and Motoki Abe},
  journal={ArXiv},
  year={2018},
  volume={abs/1804.00097}
}