[retriever] add option to use multimodal vl embedder

[edknv]

Description

• Add a LlamaNemotronEmbedVL1BV2Embedder for the nvidia/llama-nemotron-embed-vl-1b-v2 model.
• Only suppports text modality for now, so reject non-text modalities (image, text_image, image_text). Will follow on with support for non-text modalities.

Usage

python -m retriever.examples.batch_pipeline \
  /path/to/pdfs \
  --embed-model-name nvidia/llama-nemotron-embed-vl-1b-v2 \

Checklist

• I am familiar with the Contributing Guidelines.
• New or existing tests cover these changes.
• The documentation is up to date with these changes.
• If adjusting docker-compose.yaml environment variables have you ensured those are mimicked in the Helm values.yaml file.