the existing code supports both NVIDIAEmbeddings communication with the NVIDIA API Catalog as well as a downloaded NVIDIA NIM.
this PR only provides test plumbing to validate the NIM support.
currently, there is no NIM CI infrastructure, so the tests must be run manually.
after setting up an embedding NIM on port 8082, use poetry run pytest tests/integration_tests/test_embeddings.py --nim-endpoint http://localhost:8082/v1 --embedding-model-id NV-Embed-QA -vv
the existing code supports both NVIDIAEmbeddings communication with the NVIDIA API Catalog as well as a downloaded NVIDIA NIM.
this PR only provides test plumbing to validate the NIM support.
currently, there is no NIM CI infrastructure, so the tests must be run manually.
after setting up an embedding NIM on port 8082, use
poetry run pytest tests/integration_tests/test_embeddings.py --nim-endpoint http://localhost:8082/v1 --embedding-model-id NV-Embed-QA -vv