hey guys! I'm trying to use Milvus vector store with hybrid retrieval and I having issues with BGEM3F embedding model:
pip install FlagEmbedding
I create my vector store:
vector_store = MilvusVectorStore(
collection_name=app_settings.milvus_collection,
dim=app_settings.vector_dim,
overwrite=app_settings.overwrite,
token=app_settings.milvus_token,
uri=app_settings.milvus_uri,
enable_sparse=app_settings.enable_sparse,
hybrid_ranker=app_settings.hybrid_ranker,
hybrid_ranker_params=app_settings.hybrid_ranker_params,
)
and then I have this in my terminal:
DEBUG:pymilvus.milvus_client.milvus_client:Created new connection using: d28ca022b4884f4ab606408820a88944
WARNING:llama_index.vector_stores.milvus.base:Sparse embedding function is not provided, using default.
INFO:datasets:PyTorch version 2.4.0 available.
CRITICAL:llama_index.vector_stores.milvus.utils:Cannot import BGEM3FlagModel from FlagEmbedding. It seems it is not installed. Please install it using:
pip install FlagEmbedding
but in fact it has already been installed