This repository contains only the vision encoder (no text tower). It is a drop-in replacement for the SigLIP 2 vision tower: identical architecture and I/O. To use it in an MLLM, keep your existing SigLIP 2 pipeline and only change the vision-tower path to this checkpoint.

Usage

Paper: SigLIP-HD by Fine-to-Coarse Supervision
Code: <a href="https://github.com/LiheYoung/SigLIP-HD" target="_blank" rel="noopener noreferrer" class="text-blue-400 hover:text-blue-300 underline underline-offset-2 transition-colors">https://github.com/LiheYoung/SigLIP-HD

python

import torch
from PIL import Image
from transformers import SiglipVisionModel, AutoImageProcessor

model = SiglipVisionModel.from_pretrained("LiheYoung/SigLIP-HD").eval()
processor = AutoImageProcessor.from_pretrained("LiheYoung/SigLIP-HD")

image = Image.open("example.jpg").convert("RGB")
inputs = processor(images=image, return_tensors="pt")

with torch.no_grad():
    features = model(**inputs, output_hidden_states=True).hidden_states[-1]  # (1, 1024, 1152)

Citation

bibtex

@inproceedings{sigliphd,
  title={SigLIP-HD by Fine-to-Coarse Supervision},
  author={Yang, Lihe and Zhao, Zhen and Zhao, Hengshuang},
  booktitle={ICLR},
  year={2026}
}

Acknowledgement

This work is built upon SigLIP 2. We sincerely thank the authors for open-sourcing their excellent vision encoder.