Nvidia nv-embed-v2 availability via nvidiaembedding

At a glance

The community members are discussing the availability of the NVIDIA NV-Embed-v2 model through the NVIDIAEmbedding library. One community member suggests that the model is likely available, and recommends checking the embed_model.available_models to see all the available models. Another community member shares a list of available models, but does not see the NV-Embed-v2 model in the list. However, they note that the model may still be available even if it's not in the list, and suggest submitting a request if needed. Another community member confirms that the model should work even if it's not in the list, and recommends checking print(embed_model.available_models) to see all the available model names.

Useful resources

DDS

Hey all, is NVIDIA NV-Embed-v2 available via NVIDIAEmbedding?

5 comments

LLogan M

Probably? You can get all available models with embed_model.available_models when using nvidia embeddings

LLogan M

(was just looking at the source code)

DDS

Yeah I only see the following:

MODEL_ENDPOINT_MAP = {
"NV-Embed-QA": "https://ai.api.nvidia.com/v1/retrieval/nvidia/",
"snowflake/arctic-embed-l": "https://integrate.api.nvidia.com/v1/",
"nvidia/nv-embed-v1": "https://integrate.api.nvidia.com/v1/",
"nvidia/nv-embedqa-mistral-7b-v2": "https://integrate.api.nvidia.com/v1/",
"nvidia/nv-embedqa-e5-v5": "https://integrate.api.nvidia.com/v1/",
"baai/bge-m3": "https://integrate.api.nvidia.com/v1/",
"nvidia/llama-3.2-nv-embedqa-1b-v1": "https://integrate.api.nvidia.com/v1/",
}

I thought I may be missing the v2 model somewhere, but can submit a request if needed.

LLogan M

Sorry, it should work even if its not in the map (at least, from my understanding of this code)

LLogan M

If you call print(embed_model.available_models) you will see all model names that will work

Add a reply

Find answers from the community

Nvidia nv-embed-v2 availability via nvidiaembedding