Louvain clustering using transformer embeddings

Hi,

I have some document nodes with title that are linked (has_kewords {weight}) to some keyword nodes. I want to cluster them by Louvain algorithm, and I added sentence transformer embeddings of title array[flot values], since the rest of my properties are strings. But when I ran it I get only one node clusters. How can I cluster this documents base on title (I want to use GDS, not scripts in other languages)

Thanks

This is my projection graph

CALL gds.graph.project(

'myGraph',

'Document',

{

has_keywords: {

orientation: 'UNDIRECTED'

}

},

{

nodeProperties: 'embedding',

relationshipProperties: 'weight'

}

)