How to provide embeddings instead of Weavite calculating embeddings using t2v-transformers? Reason for doing that is to avoid storing content in weavite?

[csplrj]

How to provide embeddings instead of Weavite calculating embeddings using t2v-transformers? Reason for doing that is to avoid storing content in weavite?

[DudaNogueira]

hi @csplrj ! Welcome to our community!

I have just answered a thread here that touches this subject, but for image embeddings. check it out:

Optimizing Weaviate for Image Embedding Search without Storing Images

Hi @nomomon ! Welcome to our community I don’t think that one can use Weaviate vectorization modules and do not store the image with it as of now. However, you can supply the vectors yourself. So when a new image comes in, you vectorize it yourself, ingest only the image metadata (filename, geolocation, etc.), it’s vectors and the S3 reference. Notice that, whenever doing a query, you will need to also vectorize the image in order to query it against your collection. Here is a tutoria…

Here is a tutorial that might help you with that:
https://weaviate.io/developers/weaviate/tutorials/custom-vectors

Let me know if this helps!

Thanks!