Description
I am using database version 1.30.1, Python 3.9 and weaviate-client version 4.14.1. I am trying to insert multi-vector embeddings into a database, following the documentation
Any additional Information
I was following this part of the documentation to insert user-generated embeddings.
But the database is filled with a 1D vector with junk values.
Code:
collection_name = "ColbertDocs"
client.collections.create(
name=collection_name,
properties=[
Property(name="text", data_type=DataType.TEXT),
Property(name="docid", data_type=DataType.TEXT),
],
vectorizer_config=[
Configure.NamedVectors.none(
name="multi_vector", # your custom vector field name
vector_index_config=Configure.VectorIndex.hnsw(
multi_vector=Configure.VectorIndex.MultiVector.multi_vector()
)
)
]
)
# Step 2: Add an object with a 2D vector (list of vectors)
collection = client.collections.get(collection_name)
# Example 2D vector: 3 tokens x 4 dimensions
colbert_embedding = [
[0.1, 0.2, 0.3, 0.4],
[0.5, 0.6, 0.7, 0.8],
[0.9, 1.0, 1.1, 1.2],
]
collection.data.insert(
properties={"text": "The cat sat on the mat.", "docid": "123"},
vector={"multi_vector": colbert_embedding}
)
client.close()
Database values
Has someone already worked with multi-vector embeddings ? If so, can you help me with this ?
