Eva02 Tiny Patch14 224.mim In22k

🚀 Model Card for eva02_tiny_patch14_224.mim_in22k

This is an EVA02 feature / representation model. It was pretrained on ImageNet-22k with masked image modeling (using EVA-CLIP as a MIM teacher) by the paper authors.

EVA-02 models are vision transformers that incorporate mean pooling, SwiGLU, Rotary Position Embeddings (ROPE), and an extra LN in MLP (for Base & Large).

⚠️ Important Note

timm checkpoints are in float32 for consistency with other models. In some cases, the original checkpoints are in float16 or bfloat16. Refer to the originals if that's your preference.

🚀 Quick Start

Image Classification

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_tiny_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)

Image Embeddings

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_tiny_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, 192) shaped tensor

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

✨ Features

• An EVA02 feature / representation model.
• Pretrained on ImageNet-22k with masked image modeling.
• Vision transformers with mean pooling, SwiGLU, Rotary Position Embeddings (ROPE), and extra LN in MLP (for Base & Large).

📚 Documentation

Model Details

Model Comparison

Explore the dataset and runtime metrics of this model in timm model results.

📄 License

This project is licensed under the MIT license.