gc_efficientnetv2_rw_t.agc_in1k 模型卡片

一种带有全局上下文注意力的 GC-EfficientNet-v2 图像分类模型。这是 timm 中的一个特定变种架构。使用以下模板在 timm 中对 ImageNet-1k 进行训练。

训练细节：

• 基于 ResNet Strikes Back 个 C 配方
• 使用 SGD（带 Nesterov 的）优化器和 AGC（自适应梯度裁剪）
• 使用余弦学习率（Cosine LR）调度和预热（warmup）

模型详情

• 模型类型：图像分类 / 特征背骨
• 模型统计：
  • 参数数量（M）：13.7
  • GMACs：1.9
  • 激活函数数量（M）：10.0
  • 图像尺寸：训练 = 224 x 224，测试 = 288 x 288
• 相关论文：
  • EfficientNetV2: Smaller Models and Faster Training: https://arxiv.org/abs/2104.00298
  • GCNet: Non-local Networks Meet Squeeze-Excitation Networks and Beyond: https://arxiv.org/abs/1904.11492
  • ResNet strikes back: An improved training procedure in timm: https://arxiv.org/abs/2110.00476
• 数据集：ImageNet-1k
• 原始来源： https://github.com/huggingface/pytorch-image-models

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('gc_efficientnetv2_rw_t.agc_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(
    'gc_efficientnetv2_rw_t.agc_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, 24, 112, 112])
    #  torch.Size([1, 40, 56, 56])
    #  torch.Size([1, 48, 28, 28])
    #  torch.Size([1, 128, 14, 14])
    #  torch.Size([1, 208, 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(
    'gc_efficientnetv2_rw_t.agc_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, 1024, 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 中，探索该模型的数据集和运行时指标 model results 。

引用

@inproceedings{tan2021efficientnetv2,
  title={Efficientnetv2: Smaller models and faster training},
  author={Tan, Mingxing and Le, Quoc},
  booktitle={International conference on machine learning},
  pages={10096--10106},
  year={2021},
  organization={PMLR}
}

@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{cao2019GCNet,
  title={GCNet: Non-local Networks Meet Squeeze-Excitation Networks and Beyond},
  author={Cao, Yue and Xu, Jiarui and Lin, Stephen and Wei, Fangyun and Hu, Han},
  journal={arXiv preprint arXiv:1904.11492},
  year={2019}
}

@inproceedings{wightman2021resnet,
  title={ResNet strikes back: An improved training procedure in timm},
  author={Wightman, Ross and Touvron, Hugo and Jegou, Herve},
  booktitle={NeurIPS 2021 Workshop on ImageNet: Past, Present, and Future}
}