Instructions to use google/siglip2-base-patch32-256 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use google/siglip2-base-patch32-256 with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("zero-shot-image-classification", model="google/siglip2-base-patch32-256") pipe( "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/hub/parrots.png", candidate_labels=["animals", "humans", "landscape"], )# Load model directly from transformers import AutoModel model = AutoModel.from_pretrained("google/siglip2-base-patch32-256", dtype="auto") - Notebooks
- Google Colab
- Kaggle
mSigLip
#3
by Chloez - opened
Why mSigLip is not accessable here? Or maybe I just don't find it, if it is my fault, please share me the instruction to use mSigLip
There is no separate m (multilingual) variant for SigLIP 2. All SigLIP 2 models support multilingual inputs and should perform similarly to mSigLIP (for matching model size and resolution), see https://arxiv.org/abs/2502.14786 Section 3.
Thanks for your response. But for series of siglip models, I mean before siglip2 is released, why msiglip is not open sourced but only siglip models are open sourced?
It was released but is probably not on HF. It is available in big_vision (see SigLIP demo colab):
https://github.com/google-research/big_vision