I am using elixir livebook to do embeddings, then use it to search my Qdrant database:

{:ok, model_info} = Bumblebee.load_model({:hf, "BAAI/bge-base-en-v1.5"})
{:ok, tokenizer} = Bumblebee.load_tokenizer({:hf, "BAAI/bge-base-en-v1.5"})
# {:ok, model_info} = Bumblebee.load_model({:hf, "sentence-transformers/all-MiniLM-L6-v2"})
# {:ok, tokenizer} = Bumblebee.load_tokenizer({:hf, "sentence-transformers/all-MiniLM-L6-v2"})

serving = Bumblebee.Text.text_embedding(model_info, tokenizer)

text = "where do i change my password?"
emb = Nx.Serving.run(serving, text)

collection_name = "demo_v1"
Qdrant.search_points(collection_name, %{vector: emb.embedding, limit: 3})

I am getting an error:

{:error,
 {Tesla.Middleware.JSON, :encode,
  %Protocol.UndefinedError{
    protocol: Jason.Encoder,
    value: #Nx.Tensor<
      f32[768]
      [-0.8432725071907043, -0.5420605540275574, ...]
    >,
    description: "Jason.Encoder protocol must always be explicitly implemented.nnIf you own the struct, you can derive the implementation specifying which fields should be encoded to JSON:nn    @derive {Jason.Encoder, only: [....]}n    defstruct ...nnIt is also possible to encode all fields, although this should be used carefully to avoid accidentally leaking private information when new fields are added:nn    @derive Jason.Encodern    defstruct ...nnFinally, if you don't own the struct you want to encode to JSON, you may use Protocol.derive/3 placed outside of any module:nn    Protocol.derive(Jason.Encoder, NameOfTheStruct, only: [...])n    Protocol.derive(Jason.Encoder, NameOfTheStruct)n"
  }}}

from the error message, it seems like i still need to convert nx.tensor embedding to some sort of JSON format, before I can use it for the search_point/2 function in Qdrant.

How can I resolve this?