Langchain only supports vector type for now, so for dimensions superior to 2000 for exemple [text-embedding-3-large is 3072). it's not possible to create a hnsw index via langchain because of the dimension limitation. I saw a pull request to support halfvec and sparsevec https://github.com//pull/226 but it wasn't merged.
so my question is, will there be official support soon?
Langchain only supports vector type for now, so for dimensions superior to 2000 for exemple [text-embedding-3-large is 3072). it's not possible to create a hnsw index via langchain because of the dimension limitation. I saw a pull request to support halfvec and sparsevec https://github.com//pull/226 but it wasn't merged.
so my question is, will there be official support soon?