You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I have stored different data in different collections based on various sources, so I have defined multiple MilvusDocumentStores. Then, how can I retrieving from multiple document_store simultaneously in the MilvusEmbeddingRetriever? Thank you for your support. @zc277584121
retriever = MilvusEmbeddingRetriever(
document_store=self.document_store, # How to retriever multiple document_store?
top_k=top_k
)
The text was updated successfully, but these errors were encountered:
@Silence-Well I think currently MilvusEmbeddingRetriever can not satisfy this requirement. Maybe the best workaround is to subclass the class MilvusEmbeddingRetriever , and then override and implement your costomized logic in it.
My current solution is to define two MilvusEmbeddingRetriever instances and add them to the pipeline. Finally, merge the two sets of results, documents, into one docs, which can indirectly solve this problem.
I have stored different data in different collections based on various sources, so I have defined multiple MilvusDocumentStores. Then, how can I retrieving from multiple document_store simultaneously in the MilvusEmbeddingRetriever? Thank you for your support. @zc277584121
The text was updated successfully, but these errors were encountered: