#initialize dimensionality reduction model and reduce embeddings
umap_model = UMAP(n_neighbors=5, min_dist=0.0, metric='cosine', random_state=42)
reduced_embeddings = umap_model.fit_transform(embeddings)