Models / nvidia / Llama Nemotron Embed VL 1B v2

Llama Nemotron Embed VL 1B v2

JSON →
embedding

A 1-billion parameter multimodal embedding model from NVIDIA, based on Llama, designed for vision-language retrieval and embedding tasks.

textimagevisionopen-weights
Specs
Context & limits
context window
max output
Lifecycle
Dates
released2025-02-01
Resources