Served on MixpeekEmbeddingVerified 2026-11-05
DINOv2 Base
facebook/dinov2-base
DINOv2 Base is a image embedding model (Apache-2.0). Served on Mixpeek, scaling to zero when idle.
A Ray Serve app in Mixpeek's inference registry hosts it. It starts at zero replicas and wakes on first use.
New opens the Designer in Studio. For an existing namespace, open it in Studio and use its Designer tab.
No demo for DINOv2 Base yet
A video, image, slideshow or playground goes here once we record one on our own data.
What it is good and bad at
- Best at
- Image-to-image similarity on visual structure (layout, near-duplicates, the same object or scene) with no text in the loop.
- Worst at
- Text queries (no text tower, so text-to-image search is impossible); semantic matches across different-looking images, where SigLIP is better.
- Task
- Embedding
- Input
- Image
- Output
- Vector
- Use cases
- Visual search, Web pages
Templates that use it
No template uses this model yet. Browse templates.
Input and output
Input
- image
- resolution
- resize shortest side 256, center crop 224
- color
- RGB
- normalize
- ImageNet mean/std
Output
- vector
- dims
- 768
- distance
- cosine
- type
- dense
How Mixpeek wires it
- Extractors
web_scraper@v1generate_structure_embeddings=true (default true); image documents only
- Feature URIs
mixpeek://web_scraper@v1/facebook__dinov2_base
- Vector indexes
facebook__dinov2_base- Retriever stages
- feature_search
- Combines with
- web_scraper: DINOv2 structure vector beside SigLIP image vectors for the same page images
How it runs
- Status
- Served on Mixpeek Ray Serve app in the inference registry (image queries); web_scraper loads facebook/dinov2-base in-process (DocTypeStructureEmbedder).
- Serve apps
facebook__dinov2_base- Runtime
- transformers AutoModel + AutoImageProcessor (Dinov2Model)
- Weights
- safetensors, no remote code
- Parameters
- 86.6 M
- Size on disk
- 346 MB
- Preprocessing
- AutoImageProcessor defaults for the checkpoint (resize, center crop 224, ImageNet normalization).
- Postprocessing
- CLS token of last_hidden_state, L2-normalized (768 dims).
- Serving
- both
- Hardware
- Runs on CPU (slower without a GPU)
Scale to zero
- Minimum replicas
- 0
- Maximum replicas
- 4
- Idle timeout
- 300 s
- Memory
- 1 GB
- CPUs
- 0.5
- GPUs
- Set at deploy time: hardware_config.gpu_fraction if hardware_config.gpu_concurrency > 0 else 0
- Cold start, node warm
- Not recorded
- Cold start, new node
- Not recorded
Measured cost
Unmeasured
No committed throughput or cost measurement for this model.
Our evaluations
Unmeasured
We have not evaluated this model on our own data yet. Publisher results are in the sources below.
Questions
- What is DINOv2 Base best at?
- Image-to-image similarity on visual structure (layout, near-duplicates, the same object or scene) with no text in the loop.
- What are the limits of DINOv2 Base?
- Text queries (no text tower, so text-to-image search is impossible); semantic matches across different-looking images, where SigLIP is better.
- Can DINOv2 Base be used commercially?
- Yes. Apache-2.0 weights and code; no use restrictions.
- How does DINOv2 Base run on Mixpeek?
- Ray Serve app in the inference registry (image queries); web_scraper loads facebook/dinov2-base in-process (DocTypeStructureEmbedder).