Public KG record
NeoMME
Recorded overview
Single bidirectional Transformer processes both text tokens and raw image patche
An evidence-backed entity from the project knowledge graph; detailed tool metadata is still being added.
Recorded details
- Benchmarks: both model sizes lie on ViDoRe v3 Pareto frontier for nDCG@10 and model size
- Capabilities: generates vector representations for input text and/or images
- Capabilities: handles visual document retrieval without OCR
- Capabilities: uses grouped-query attention, query-key normalization, gated attention, 2D rotary position embeddings, and squared-ReLU MLPs
- Deployment: cloud/local
- Integrations: Hugging Face Transformers
- Integrations: Transformers
- Languages: multilingual
Evidence coverage
The KG currently records 14 known information dimensions and has 4 undocumented dimensions, linked to 2 public source articles. These counts describe record coverage, not model performance or verified fit.
Related public records
- H Company — releases
- ModernBERT — related to
- NeoMME-Retriever — based on
- Transformers — integrates
- Bidirectional Encoder — is a