You can deploy this model and start earning money today!

google/
siglip2-base-patch16-naflex

Name: siglip2-base-patch16-naflex
Rating: 2.5 (1224479 reviews)

transformers

safetensors

siglip2

zero-shot-image-classification

Views

No views yet

SigLIP 2 Base

SigLIP 2 extends the pretraining objective of SigLIP with prior, independently developed techniques into a unified recipe, for improved semantic understanding, localization, and dense features.

Intended uses

You can use the raw model for tasks like zero-shot image classification and image-text retrieval, or as a vision encoder for VLMs (and other vision tasks).

Here is how to use this model to perform zero-shot image classification:

python

from transformers import pipeline

# load pipeline
ckpt = "google/siglip2-base-patch16-naflex"
image_classifier = pipeline(model=ckpt, task="zero-shot-image-classification")

# load image and candidate labels
url = "http://images.cocodataset.org/val2017/000000039769.jpg"
candidate_labels = ["2 cats", "a plane", "a remote"]

# run inference
outputs = image_classifier(image, candidate_labels)
print(outputs)

You can encode an image using the Vision Tower like so:

python

import torch
from transformers import AutoModel, AutoProcessor
from transformers.image_utils import load_image

# load the model and processor
ckpt = "google/siglip2-base-patch16-naflex"
model = AutoModel.from_pretrained(ckpt, device_map="auto").eval()
processor = AutoProcessor.from_pretrained(ckpt)

# load the image
image = load_image("https://huggingface.co/datasets/merve/coco/resolve/main/val2017/000000000285.jpg")
inputs = processor(images=[image], return_tensors="pt").to(model.device)

# run infernece
with torch.no_grad():
    image_embeddings = model.get_image_features(**inputs)    

print(image_embeddings.shape)

For more code examples, we refer to the siglip2 documentation.

Training procedure

SigLIP 2 adds some clever training objectives on top of SigLIP:

Decoder loss
Global-local and masked prediction loss
Aspect ratio and resolution adaptibility

Training data

SigLIP 2 is pre-trained on the WebLI dataset (Chen et al., 2023).

Compute

The model was trained on up to 2048 TPU-v5e chips.

Evaluation results

Evaluation of SigLIP 2 is shown below (taken from the paper).

BibTeX entry and citation info

bibtex

@misc{tschannen2025siglip2multilingualvisionlanguage,
      title={SigLIP 2: Multilingual Vision-Language Encoders with Improved Semantic Understanding, Localization, and Dense Features}, 
      author={Michael Tschannen and Alexey Gritsenko and Xiao Wang and Muhammad Ferjad Naeem and Ibrahim Alabdulmohsin and Nikhil Parthasarathy and Talfan Evans and Lucas Beyer and Ye Xia and Basil Mustafa and Olivier Hénaff and Jeremiah Harmsen and Andreas Steiner and Xiaohua Zhai},
      year={2025},
      eprint={2502.14786},
      archivePrefix={arXiv},
      primaryClass={cs.CV},
      url={https://arxiv.org/abs/2502.14786}, 
}