eva02_large_patch14_224.mim_in22k的模型卡片

一个EVA02特征/表示模型。由论文作者使用EVA-CLIP作为MIM教师在ImageNet-22k上进行了预训练，并使用遮蔽图像建模。

EVA-02模型是视觉变换器，具有平均池化、SwiGLU、Rotary Position Embeddings（ROPE）和MLP中额外的LN（适用于Base和Large）。

注意：timm检查点为float32，以保持与其他模型的一致性。原始检查点在某些情况下为float16或bfloat16，如果需要，可以查看原始检查点。

模型细节

• 模型类型：图像分类/特征主干
• 模型统计：
  • 参数（M）：303.3
  • GMACs：81.1
  • 激活值（M）：97.2
  • 图像尺寸：224 x 224
• 论文：
  • EVA-02：Neon Genesis的视觉表示： https://arxiv.org/abs/2303.11331
  • EVA-CLIP：CLIP的改进训练技术： https://arxiv.org/abs/2303.15389
• 原始模型：
  • https://github.com/baaivision/EVA
  • https://huggingface.co/Yuxin-CV/EVA-02
• 预训练数据集：ImageNet-22k

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('eva02_large_patch14_224.mim_in22k', 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)

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(
    'eva02_large_patch14_224.mim_in22k',
    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, 257, 1024) 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{EVA02,
  title={EVA-02: A Visual Representation for Neon Genesis},
  author={Fang, Yuxin and Sun, Quan and Wang, Xinggang and Huang, Tiejun and Wang, Xinlong and Cao, Yue},
  journal={arXiv preprint arXiv:2303.11331},
  year={2023}
}

@article{EVA-CLIP,
  title={EVA-02: A Visual Representation for Neon Genesis},
  author={Sun, Quan and Fang, Yuxin and Wu, Ledell and Wang, Xinlong and Cao, Yue},
  journal={arXiv preprint arXiv:2303.15389},
  year={2023}
}

@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}}
}