Vit B 16 SigLIP I18n 256

🚀 ViT-B-16-SigLIP-i18n-256 Model Card

A SigLIP (Sigmoid loss for Language-Image Pre-training) model trained on WebLI, enabling zero-shot image classification.

🚀 Quick Start

This SigLIP model is trained on WebLI and has been converted from the original JAX checkpoints in Big Vision to PyTorch. The weights can be used in both OpenCLIP (image + text) and timm (image only).

✨ Features

• Contrastive Image-Text: Capable of handling both image and text data for contrastive learning.
• Zero-Shot Image Classification: Can perform image classification without explicit fine - tuning on specific datasets.

📦 Installation

No specific installation steps are provided in the original document.

💻 Usage Examples

Basic Usage

With OpenCLIP

import torch
import torch.nn.functional as F
from urllib.request import urlopen
from PIL import Image
from open_clip import create_model_from_pretrained, get_tokenizer # works on open-clip-torch>=2.23.0, timm>=0.9.8

model, preprocess = create_model_from_pretrained('hf-hub:timm/ViT-B-16-SigLIP-i18n-256')
tokenizer = get_tokenizer('hf-hub:timm/ViT-B-16-SigLIP-i18n-256')

image = Image.open(urlopen(
    'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png'
))
image = preprocess(image).unsqueeze(0)

labels_list = ["a dog", "a cat", "a donut", "a beignet"]
text = tokenizer(labels_list, context_length=model.context_length)

with torch.no_grad(), torch.cuda.amp.autocast():
    image_features = model.encode_image(image)
    text_features = model.encode_text(text)
    image_features = F.normalize(image_features, dim=-1)
    text_features = F.normalize(text_features, dim=-1)

    text_probs = torch.sigmoid(image_features @ text_features.T * model.logit_scale.exp() + model.logit_bias)

zipped_list = list(zip(labels_list, [round(p.item(), 3) for p in text_probs[0]]))
print("Label probabilities: ", zipped_list)

With timm (for image embeddings)

from urllib.request import urlopen
from PIL import Image
import timm

image = Image.open(urlopen(
    'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png'
))

model = timm.create_model(
    'vit_base_patch16_siglip_256',
    pretrained=True,
    num_classes=0,
)
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(image).unsqueeze(0))  # output is (batch_size, num_features) shaped tensor

📚 Documentation

Model Details

📄 License

This model is licensed under the apache - 2.0 license.

🔧 Technical Details

This model is a SigLIP model trained on the WebLI dataset. It uses a sigmoid loss function for language - image pre - training. The model has been converted from the original JAX checkpoints in Big Vision to PyTorch, and the weights are compatible with both OpenCLIP and timm.

📖 Citation

@article{zhai2023sigmoid,
  title={Sigmoid loss for language image pre-training},
  author={Zhai, Xiaohua and Mustafa, Basil and Kolesnikov, Alexander and Beyer, Lucas},
  journal={arXiv preprint arXiv:2303.15343},
  year={2023}
}

@misc{big_vision,
  author = {Beyer, Lucas and Zhai, Xiaohua and Kolesnikov, Alexander},
  title = {Big Vision},
  year = {2022},
  publisher = {GitHub},
  journal = {GitHub repository},
  howpublished = {\url{https://github.com/google-research/big_vision}}
}