New posts in Gmm

Which algorithm and what combination of hyper-parameters will be the best to cluster this data?