siglip2-giant-opt-patch16-384

Name: siglip2-giant-opt-patch16-384
Author: google

by google

Multilingual vision-language encoder with dense features and localization

2.6Mdl/month

43likes

1.9Bparams

HuggingFace Run it on your own data, free

Identifiers

Model ID

google/siglip2-giant-opt-patch16-384

Feature URI

mixpeek://image_extractor@v1/google_siglip2_giant_v1

Overview

SigLIP 2 extends the sigmoid contrastive objective with captioning-based pretraining, self-supervised losses, and online data curation into a unified recipe. It produces stronger vision-language encoders with significantly improved localization and dense feature quality.

On Mixpeek, SigLIP 2 provides the strongest zero-shot visual embeddings from Google, achieving 85.0% ImageNet accuracy at the giant scale. Its improved spatial understanding makes it ideal for tasks requiring localization alongside retrieval.

Architecture

Vision Transformer (ViT-g) with ~1B parameters at 384px resolution. Combines sigmoid contrastive loss with captioning, self-distillation, and masked prediction objectives. Supports multi-resolution and native aspect ratio inputs.

Mixpeek SDK Integration

// No extractor parameter takes a Hugging Face model id (checked against
// GET /v1/discovery/extractors, which returns 13), so siglip2-giant-opt-patch16-384 runs
// on your side and the output is upserted through POST
// /v1/namespaces/{namespace_id}/documents/upsert. On Enterprise the other
// path is to upload the weights instead: POST /v1/namespaces/{id}/models
// accepts the huggingface format and a custom plugin loads them.
const res = await fetch(
  "https://api.mixpeek.com/v1/namespaces/ns_your_namespace/documents/upsert",
  {
    method: "POST",
    headers: {
      Authorization: "Bearer API_KEY",
      "Content-Type": "application/json",
    },
    body: JSON.stringify({
      collection_id: "col_your_collection",
      documents: [
        {
          document_id: "asset-00412",
          // The vector name has to match a vector index on the collection.
          vectors: { "image-embedding": yourVector },
          payload: { source_key: "archive/2026/asset-00412" },
        },
      ],
    }),
  },
);

// Managed alternative, if this exact model is not the requirement:
// image_extractor@v1 runs google/siglip-base-patch16-224
// (768-d) over a bucket, with no inference of your own.