Production hybrid retrieval using bgem3

[swapnilashtekar]

Hi, What are the model inferencing options to use BGE-M3 model in production scenario? I am using model to obtain hybrid retrieval (both dense and sparse embeddings) as mentioned in milvus documentation. https://milvus.io/api-reference/pymilvus/v2.5.x/EmbeddingModels/BGEM3EmbeddingFunction/BGEM3EmbeddingFunction.md Are there any ways where i can run bge-m3 model to get hybrid retrieval as mentioned in https://github.com/huggingface/text-embeddings-inference Thank you.

[hanhainebula]

Hello, @swapnilashtekar. I'm sorry that I'm not familiar with text-embeddings-inference. You can try opening an issue at their repo.

However, we provide some examples here about how to perform hybrid retrieval with BGE-M3. As you said, milvus also supports hybrid retrieval with BGE-M3.