UltraVBERT
Collection
Tiny vision-language models for visual document retrieval, built on UltraBERT. Coming soon! • 5 items • Updated
BiUltraVBERT
A single-vector retriever for visual documents: UltraVBERT with a mean-pooled 768-d embedding per page and query.
Coming soon!
Models
| Model | Role | Status |
|---|---|---|
| UltraVBERT | Base encoder aligned to page images | Available |
| BiUltraVBERT | Single-vector retriever for visual documents | Cooking |
| ColUltraVBERT | Late-interaction retriever for visual documents | Available · Demo |
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support