add support for local embeddings with hf transformers

[furlat]

https://huggingface.co/jinaai/jina-embeddings-v2-base-en https://huggingface.co/nomic-ai/nomic-embed-text-v1

[pszemraj]

here's a gist I made covering usage of the nomic-text-v1 model with ONNX - might help later.

there are likely further optimizations within onnx (or even openvino) to be made but I can get 2ish docs/sec at 8192 ctx on CPU, could be worse