NVIDIA's generalist embedding model achieving record 69.32 score on MTEB benchmark. Fine-tuned from Llama architecture with improved techniques for training LLMs as embedding models.
NV-Embed is a generalist embedding model that enhances the performance of decoder-only LLMs for embedding and retrieval tasks, with various architectural designs and training procedures.
Using only publicly available data, NV-Embed achieved:
NVIDIA has developed several embedding models based on Llama architecture:
Available through:
Loading more......