mobilevitv2_175.cvnets_in22k_ft_in1k_384的模型卡片

一个MobileViT-v2图像分类模型。由论文作者在ImageNet-22k上预训练，并在ImageNet-1k上进行微调。

https://github.com/apple/ml-cvnets/blob/main/LICENSE 处查看许可证详细信息

模型详情

• 模型类型：图像分类 / 特征主干
• 模型统计信息:
  • 参数数量 (M)：14.3
  • GMACs：12.5
  • 激活数量 (M)：63.3
  • 图像尺寸：384 x 384
• 论文：
  • 移动视觉变换器的可分离自注意力机制： https://arxiv.org/abs/2206.02680
• 原始链接： https://github.com/apple/ml-cvnets
• 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('mobilevitv2_175.cvnets_in22k_ft_in1k_384', 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(
    'mobilevitv2_175.cvnets_in22k_ft_in1k_384',
    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, 112, 192, 192])
    #  torch.Size([1, 224, 96, 96])
    #  torch.Size([1, 448, 48, 48])
    #  torch.Size([1, 672, 24, 24])
    #  torch.Size([1, 896, 12, 12])

    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(
    'mobilevitv2_175.cvnets_in22k_ft_in1k_384',
    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, 896, 12, 12) shaped tensor

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

模型比较

在timm model results 中探索该模型的数据集和运行时指标。

引用

@article{Mehta2022SeparableSF,
  title={Separable Self-attention for Mobile Vision Transformers},
  author={Sachin Mehta and Mohammad Rastegari},
  journal={ArXiv},
  year={2022},
  volume={abs/2206.02680}
}