许可协议: cc-by-nc-4.0
库名称: timm
标签:

• 图像特征提取
• timm
• transformers

convnextv2_atto.fcmae 模型卡

一个基于ConvNeXt-V2的自监督特征表示模型。采用全卷积掩码自编码器框架（FCMAE）进行预训练。该模型没有预训练头部，仅适用于微调或特征提取。

模型详情

• 模型类型: 图像分类/特征骨干网络
• 模型统计:
  • 参数量（百万）: 3.4
  • GMACs运算量: 0.6
  • 激活值（百万）: 3.8
  • 图像尺寸: 224 x 224
• 相关论文:
  • ConvNeXt V2: 与掩码自编码器协同设计与扩展的卷积网络: https://arxiv.org/abs/2301.00808
• 原始实现: https://github.com/facebookresearch/ConvNeXt-V2
• 预训练数据集: ImageNet-1k

模型使用

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('convnextv2_atto.fcmae', pretrained=True)  
model = model.eval()  

# 获取模型特定的变换（归一化、调整尺寸）  
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))  # 将单张图像扩展为批次大小为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(  
    'convnextv2_atto.fcmae',  
    pretrained=True,  
    features_only=True,  
)  
model = model.eval()  

# 获取模型特定的变换（归一化、调整尺寸）  
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))  # 将单张图像扩展为批次大小为1  

for o in output:  
    # 打印输出中每个特征图的形状  
    # 例如:  
    #  torch.Size([1, 40, 56, 56])  
    #  torch.Size([1, 80, 28, 28])  
    #  torch.Size([1, 160, 14, 14])  
    #  torch.Size([1, 320, 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(  
    'convnextv2_atto.fcmae',  
    pretrained=True,  
    num_classes=0,  # 移除分类器nn.Linear  
)  
model = model.eval()  

# 获取模型特定的变换（归一化、调整尺寸）  
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))  # 输出为（批次大小, 特征数量）形状的张量  

# 或等效方式（无需设置num_classes=0）  

output = model.forward_features(transforms(img).unsqueeze(0))  
# 输出为未池化的（1, 320, 7, 7）形状张量  

output = model.forward_head(output, pre_logits=True)  
# 输出为（1, 特征数量）形状的张量  

模型比较

在timm的模型结果中探索该模型的数据集和运行时指标。

所有计时数据均来自RTX 3090上使用AMP的PyTorch 1.13的eager模式。

引用

@article{Woo2023ConvNeXtV2,  
  title={ConvNeXt V2: Co-designing and Scaling ConvNets with Masked Autoencoders},  
  author={Sanghyun Woo, Shoubhik Debnath, Ronghang Hu, Xinlei Chen, Zhuang Liu, In So Kweon and Saining Xie},  
  year={2023},  
  journal={arXiv preprint arXiv:2301.00808},  
}  

@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}}  
}