metadata

tags:
  - clip
  - siglip
library_name: open_clip
pipeline_tag: zero-shot-image-classification
license: apache-2.0
datasets:
  - webli

Model card for ViT-B-16-SigLIP-512

A SigLIP (Sigmoid loss for Language-Image Pre-training) model trained on WebLI.

This model has been converted to PyTorch from the original JAX checkpoints in Big Vision. These weights are usable in both OpenCLIP (image + text) and timm (image only).

Model Details

Model Type: Contrastive Image-Text, Zero-Shot Image Classification.
Original: https://github.com/google-research/big_vision
Dataset: WebLI
Papers:
- Sigmoid loss for language image pre-training: https://arxiv.org/abs/2303.15343

Model 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

model, preprocess = create_model_from_pretrained('hf-hub:ViT-B-16-SigLIP-512')
tokenizer = get_tokenizer('hf-hub:ViT-B-16-SigLIP-512')

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

text = tokenizer(["a diagram", "a dog", "a cat", "a beignet"], 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 = (100.0 * image_features @ text_features.T).softmax(dim=-1)

print("Label probs:", text_probs)  # prints: [[0., 0., 0., 1.0]]

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_512',
    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

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}}
}