---
title: google/siglip-so400m-patch14-224
description: SigLIP model pre-trained on WebLi at resolution 224x224. It was introduced in the paper Sigmoid Loss for Language Image Pre-Training by Zhai. SigLIP, 877M parameters.
canonical_url: https://superlinked.com/models/google-siglip-so400m-patch14-224
last_updated: 2026-05-25
---

# google/siglip-so400m-patch14-224

SigLIP model pre-trained on WebLi at resolution 224x224. It was introduced in the paper Sigmoid Loss for Language Image Pre-Training by Zhai et al. and first released in this repository.

Source: [google/siglip-so400m-patch14-224 on HuggingFace](https://huggingface.co/google/siglip-so400m-patch14-224)

## Overview

| Field | Value |
|-------|-------|
| Architecture | SigLIP |
| Parameters | 877M |
| Tasks | Encode |
| Outputs | Dense |
| Dimensions | Dense: 1,152 |
| Max sequence length | 64 tokens |
| License | apache-2.0 |
| Inputs | text, image |

## Benchmarks

### Flickr30kI2TRetrieval

Domain: general · Task: retrieval · Language: en

Image-to-text retrieval: retrieve captions from images

Corpus: 31,783 · Queries: 1,000

**Quality:** ndcg at 10: 0.8382 · map at 10: 0.7479 · mrr at 10: 0.9353

**Performance (L4-SPOT b1 c8):** Corpus 223 tok/s · Corpus p50 395.0ms · Query 11.5 img/s · Query p50 392.1ms

**Performance (L4 b1 c16):** Corpus 577 tok/s · Corpus p50 393.2ms · Query 7.0 mpix/s · Query p50 372.1ms

[Reference](https://shannon.cs.illinois.edu/DenotationGraph/)