Deliver precise visual search for highly specific queries. hyper3-clip uses hyperbolic geometry to understand complex hierarchies and rank the exact matches standard models miss.
Each slice tests whether hyper3-clip ranks exact visual matches higher when hierarchy, variants, or compositional details matter.
Retrieve the exact product variant based on fine-grained visual details.
grey velvet tufted sofa, brass legs, mid-century
Mid-Century Modern, Grey Velvet, Brass Legs
Rivet Freemont Sofa
Rank #1 • Sofa
Rivet Uptown Sofa
Rank #2 • Sofa
Rivet Freemont Sofa
Rank #3 • Sofa
Rivet Freemont Sofa
Rank #4 • Sofa
Rivet Uptown Sofa
Rank #5 • Sofa
Rivet Uptown Sofa
Rank #1 • Sofa
Rivet Alden Sofa
Rank #2 • Sofa
Rivet Uptown Sofa
Rank #3 • Sofa
Rivet Bradford Sofa
Rank #4 • Sofa
Rivet Accent Chair
Rank #5 • Chair
Leaf nodes crowd tightly near the bottom. Sibling details overlap and cause retrieval confusion.
We combine specialized vision-language models with an open-source inspection workbench, then test them against concrete retrieval failure modes by industry.
| Industry / Dataset | Benchmark | hyper3-CLIP | OpenAI-CLIP | Readout |
|---|---|---|---|---|
| Ecommerce Catalog Retrieval – Amazon Berkeley Objects | ||||
| Retail catalogs500 product images, 20 product types | Product-type mAP | 0.582 | 0.552 | +3.05 pts |
| Retail catalogs50 parsed catalog departments | Department mAP | 0.264 | 0.212 | +5.20 pts |
| Retail catalogsParent category retrieves diverse children | Child coverage@50 | 0.780 | 0.655 | +12.50 pts |
| Fashion Product Search – DeepFashion In-Shop | ||||
| Apparel retailSame-item product image retrieval | mAP | 0.407 | 0.240 | +16.7 pts |
| Apparel retailSame-item first-result recovery | Recall@1 | 0.595 | 0.375 | +22.0 pts |
| Apparel retailSpecific typed product search | Hit@10 | 0.572 | 0.550 | +2.2 pts |
| General Visual Hierarchy – COCO Objects | ||||
| Object search5,000 COCO val images, 80 categories | Category mAP | 0.554 | 0.532 | +2.22 pts |
| Object search12 object supercategories | Supercategory mAP | 0.536 | 0.516 | +2.08 pts |
| Object searchCoverage of child types under broad labels | Child coverage@100 | 0.887 | 0.951 | CLIP +6.40 pts |
A specialized vision-language model trained with hierarchy-aware geometry to reduce category and sibling bleed in dense visual catalogs.
An open-source viewer for mapping embedding projections, tracing nearest neighbors, and identifying the failure modes behind bad retrieval results.
Understand how our frontier specialized models are trained from scratch, and explore our latest preprints and technical deep-dives.
Why mainstream ML's reliance on Euclidean manifolds is a mistake, and how hyperbolic spaces can efficiently encode hierarchical data with fewer dimensions.
read post →Send a representative retrieval slice. We compare hyper3-clip with your current baseline and return metrics, ranked examples, and a pilot recommendation.
Send us a small sample of your highly specific queries. Provide a handful of images, expected matches, and your current baseline.
We benchmark hyper3-clip against your baseline and inspect the missed exact matches.
Receive metrics, ranked examples, and a clear recommendation on whether the model is a fit.