Clustering via Kmeans, Agglomorative, or DBSCAN/HDBSCAN.

• Kmeans. sklearn.cluster.KMeans, Faiss.kmeans. 
• Finding optimal number of clusters n_clusters. kmeans.inertia_ inside a loop, pass results to kneed.KneeLocator (Gnothi sample code). Or silhouette_score  (preferred).
• Agglomorative clustering with pre-computed cosine similarity square matrix (Gnothi sample code)
• DBSCAN / HDBSCAN