NEWVectors or files. Pick a path.Start →
    Models/Image Feature Extraction/facebook/dinov3-vits16-pretrain-lvd1689m
    Image Feature Extractiontransformersother

    dinov3-vits16-pretrain-lvd1689m

    by facebook

    DINOv3 at 21.6M parameters, small enough to run over everything

    Identifier
    Model ID
    facebook/dinov3-vits16-pretrain-lvd1689m

    Overview

    The smallest DINOv3 checkpoint, distilled from the same 7B teacher as the larger ones. Self-supervised, no text tower, so it groups images by how they look rather than by what a caption would say about them.

    Size is the whole argument here. At 21.6M parameters this runs over an entire image library on hardware that would choke on the ViT-B, which makes it the right first pass for deduplication and near-duplicate detection where you need to touch every file rather than a sample.

    Use a larger checkpoint when precision on hard pairs matters more than covering the corpus.

    Architecture

    Vision transformer, patch size 16, DINOv3ViTModel with 21,596,544 parameters. Self-supervised training on LVD-1689M, distilled from the 7B teacher. Image-feature-extraction only: no text encoder, so it cannot answer a text query on its own.

    Key Capabilities

    • Dense visual features at a size that scales to whole archives
    • Near-duplicate detection without labels
    • Frame-level features for clustering unlabelled footage
    • A cheap first pass ahead of a larger visual encoder

    Use Cases on Mixpeek

    • Deduplicating a large image library where a bigger model is unaffordable per file
    • Reverse image search over object storage
    • Clustering an archive before anyone has written a taxonomy
    • Screening candidates for a more expensive encoder to rerank

    Tags

    transformerssafetensorsdinov3_vitimage-feature-extractiondinodinov3arxiv:2508.10104enbase_model:facebook/dinov3-vit7b16-pretrain-lvd1689mbase_model:finetune:facebook/dinov3-vit7b16-pretrain-lvd1689mlicense:otherendpoints_compatibleregion:us

    Use dinov3-vits16-pretrain-lvd1689m on Mixpeek

    Build multimodal processing pipelines with this model and others. Extract features, run inference, and set up retrieval in Mixpeek Studio.

    Open Studio

    How It Runs on Mixpeek

    On Mixpeek, dinov3-vits16-pretrain-lvd1689m runs as a managed extractor inside a processing pipeline. Point a bucket of image feature extraction data at it, and Mixpeek handles GPU provisioning, batching, retries, and writing the outputs into a vector store you can query.

    Extractor outputs land in the Mixpeek Vector Store (MVS), where you can combine them with retrieval, reranking, and filter stages to build end-to-end search and agent-perception pipelines, no model-serving infrastructure to maintain.