316 1 week ago

Nemotron-3.5-Embed 1B and 8B: multilingual text + image embedding models for retrieval and RAG. 32,768 tokens context, 34 languages, bundled vision projector.

vision embedding 1b 8b

8 models