train_hdp.py

hdp_model = HDPModel(min_df=min_df, rm_top=rm_top)
hdp_model.optim_interval = 5
for d in docs_train:
    hdp_model.add_doc(d)
hdp_model.burn_in = 100
hdp_model.train(0, workers=workers)
for i in range(0, 1000, 10):
    hdp_model.train(10, workers=workers)
    print('Iteration: {}\tLog-likelihood: {}\tNum. of topics: {}'.format(i, hdp_model.ll_per_word, hdp_model.live_k))

num_of_topics = hdp_model.live_k