How to change the dimensions output using llama-3.2-nv-embedqa-1b-v2

I would like to use llama-3.2-nv-embedqa-1b-v2 to integrate with pgvector, but the 2048 output is too large for pgvector. The documentation says the output can be configured to a different value (i.e., 1024). Is this a value that gets passed in the input, or is it something entirely different?

Output: Model outputs embedding vectors of maximum dimension 2048 for each text string (can be configured based on 384, 512, 768, 1024, or 2048).

Hi @crslen welcome to the NVIDIA Developer forums!

You should now be able to use the dimensions API parameter to specify the dimension Reference - NVIDIA Docs for the llama-2.3-nv-embedqa-1b-v2 model.
Please let us know if this solves your problem.

Thanks!

Sophie