Runs as a custom pluginEmbeddingVerified 2026-11-12
MetaCLIP ViT-H/14 (2.5B)
facebook/metaclip-h14-fullcc2.5b
MetaCLIP ViT-H/14 (2.5B) is a image and text embedding model (cc-by-nc-4.0, non-commercial). Can run on Mixpeek as an imported Hugging Face model.
Mixpeek does not run it today. It can run as a custom plugin: the weights load without remote code.
No demo for MetaCLIP ViT-H/14 (2.5B) 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
- Published COCO 5k recall@1 57.5 and Flickr30k 85.0.
- Worst at
- Non-commercial license; measured 6.5 points under its paper on Flickr30k through Mixpeek, undiagnosed (SOTA bench finding 13).
- Task
- Embedding
- Input
- Image, Text
- Output
- Vector
- Use cases
- Visual search
Templates that use it
No template uses this model yet. Browse templates.
Input and output
Input
Image
- Resolution
- 224x224
- Color
- RGB
- Normalize
- CLIP mean (0.4815, 0.4578, 0.4082), std (0.2686, 0.2613, 0.2758)
Text
- Max tokens
- 77
Prompts
captions as written; CLIP BPE tokenizer (lowercases), 77 tokens
Output
Vector
- Dims
- 1024
- Distance
- cosine
- Type
- dense
Note
image and text vectors share one space; L2-normalize before cosine
How Mixpeek wires it
- Extractors
- None today
- Feature URIs
- None today
- Vector indexes
- None
- Retriever stages
- feature_search, mmr
- Combines with
- video frames (VideoFrameExpander) -> image vectors -> feature_search with a text query
- detector boxes (RT-DETR, D-FINE) -> crops -> image vectors per object
- image vectors + E5 text vectors on captions -> feature_search fusion (dbsf)
How it runs
- Status
- Runs as a custom plugin Imported from the Hugging Face Hub as an hf:// extractor and served by the transformers image-text loader, safetensors only, trust_remote_code false. GPU recommended at this size.
- Serve apps
- None
- Runtime
- transformers CLIPModel + CLIPProcessor
- Weights
- safetensors, no remote code
- Parameters
- 986 M
- Size on disk
- 3,945 MB
- Preprocessing
- Resize and center-crop to 224, CLIP mean/std normalize; tokenize to 77.
- Postprocessing
- L2-normalize; cosine similarity.
- Serving
- both, batch 32
- Hardware
- Runs on CPU (slower without a GPU)
Scale to zero
- Minimum replicas
- 0
- Maximum replicas
- 1
- Idle timeout
- 600 s
- Memory
- 6 GB
- CPUs
- 2
- GPUs
- 1
- Cold start, node warm
- Not recorded
- Cold start, new node
- Not recorded
Measured cost
Unmeasured
No committed measurement on Mixpeek yet; the SOTA bench measures it.
Our evaluations
Unmeasured
We have not evaluated this model on our own data yet. Publisher results are in the sources below.
Questions
- What is MetaCLIP ViT-H/14 (2.5B) best at?
- Published COCO 5k recall@1 57.5 and Flickr30k 85.0.
- What are the limits of MetaCLIP ViT-H/14 (2.5B)?
- Non-commercial license; measured 6.5 points under its paper on Flickr30k through Mixpeek, undiagnosed (SOTA bench finding 13).
- Can MetaCLIP ViT-H/14 (2.5B) be used commercially?
- No. CC BY-NC 4.0: non-commercial use only.
- How does MetaCLIP ViT-H/14 (2.5B) run on Mixpeek?
- Imported from the Hugging Face Hub as an hf:// extractor and served by the transformers image-text loader, safetensors only, trust_remote_code false. GPU recommended at this size.
- What does MetaCLIP ViT-H/14 (2.5B) output?
- A 1024-dimensional dense vector, compared by cosine similarity.