Is there a way to find the optimal number of clusters to use for raster map?
I considered calculating the silhouette score since the required inputs would be model.Usv (data points) and model.embedding_clust( (labels), however using the method on the dataset in rastermap_largescale.ipynb yielded a silhouette score close to 0.