Loading embeddings from a .gguf file requires llama-cpp and only works in a docker environment. This works when using faiss-search as the main repo but does not work when calling faiss-search as an import from another repo.
One fix could be to have docker integrated into the packaging to support this.
Another workaround could be to setup the dockerfile of the library it's being imported into so that it knows to grab llama-cpp if it sees faiss-search as an import.
Loading embeddings from a
.gguffile requiresllama-cppand only works in a docker environment. This works when usingfaiss-searchas the main repo but does not work when callingfaiss-searchas an import from another repo.One fix could be to have docker integrated into the packaging to support this.
Another workaround could be to setup the dockerfile of the library it's being imported into so that it knows to grab
llama-cppif it seesfaiss-searchas an import.