Instructions to use google/siglip2-base-patch16-224 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use google/siglip2-base-patch16-224 with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("zero-shot-image-classification", model="google/siglip2-base-patch16-224") 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-patch16-224", dtype="auto") - Notebooks
- Google Colab
- Kaggle
The accuracy on the ImageNet dataset is low
#13
by qingshuiL - opened
I used clip_benchmark to evaluate the model weights, and the accuracy on imagNet-1K is only 69.8. Is there anything to note here?
Hi, I'm struggling with the similar issue, and I posted a new post at: https://discuss.huggingface.co/t/siglip-2-models-show-lower-zero-shot-accuracy-than-reported/166735
Did you resolve this issue?