Module nvembed_v2 

NV-Embed-v2

NV-Embed-v2 is a text embedding model that combines a Mistral decoder with a latent attention mechanism to produce high-quality text embeddings.

This implementation is based on the paper and weights

Query-Passage Retrieval Example

cargo run --example nvembed_v2 --release

Sentence Embedding Example

cargo run --example nvembed_v2 --release -- --prompt "Here is a test sentence"

Modules

embedding

model