NEWVectors or files. Pick a path.Start →
    Models/Embeddings/facebook/dinov2-large
    HFVisual Embeddingsapache-2.0

    dinov2-large

    by facebook

    Self-supervised vision foundation model producing all-purpose visual features

    799Kdl/month
    117likes
    304Mparams
    Identifiers
    Model ID
    facebook/dinov2-large
    Feature URI
    mixpeek://image_extractor@v1/facebook_dinov2_large_v1

    Overview

    DINOv2 is a self-supervised vision foundation model from Meta AI that learns robust visual features without any labels. Trained on a curated dataset of 142M images (LVD-142M) using a combination of DINO and iBOT objectives, it produces dense features that work across image distributions and tasks without fine-tuning.

    On Mixpeek, DINOv2 provides high-quality visual embeddings for similarity search, classification, and dense prediction tasks. Its features are especially strong for fine-grained visual understanding.

    Architecture

    Vision Transformer (ViT-L/14) with 24 layers, 1024-dim hidden size, 16 attention heads. Trained via self-distillation from a 1B-parameter ViT-g teacher. Includes register tokens to fix attention artifacts in feature maps.

    Mixpeek SDK Integration

    // No extractor parameter takes a Hugging Face model id (checked against
    // GET /v1/discovery/extractors, which returns 13), so dinov2-large 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.

    Capabilities

    • Self-supervised visual features without any labels
    • 1024-dimensional dense embeddings per patch
    • Linear-probe classification at 87.1% ImageNet accuracy (ViT-g)
    • Strong on depth estimation, segmentation, retrieval
    • Register tokens for clean dense feature maps

    Use Cases on Mixpeek

    Visual similarity search across image and video libraries
    Fine-grained product matching and deduplication
    Dense feature extraction for segmentation and depth estimation
    Domain-agnostic visual representation for downstream models

    Benchmarks

    DatasetMetricScoreSource
    ImageNet (linear probe)Top-1 Accuracy81.6%Oquab et al., 2024: Table 1
    ADE20k (linear seg.)mIoU49.0Oquab et al., 2024: Table 3

    Performance

    Input Size224×224 px
    Embedding Dim1024
    GPU Latency~10ms / image (A100)
    CPU Latency~120ms / image
    GPU Throughput~100 images/sec (A100)
    GPU Memory~1.2 GB

    Specification

    FrameworkHF
    Organizationfacebook
    FeatureVisual Embeddings
    Output768-dim vector
    Modalitiesvideo, image
    RetrieverVector Search
    Parameters304M
    Licenseapache-2.0
    Downloads/mo799K
    Likes117

    Research Paper

    DINOv2: Learning Robust Visual Features without Supervision

    arxiv.org

    Build a pipeline with dinov2-large

    Add this model to a processing pipeline alongside other extractors. Combine with retrieval stages for end-to-end search.

    Run it on your own data, free