tree_stats

aligner.multiprocessing.tree_stats(directory, align_directory, split_directory, ci_phones, num_jobs, fmllr=False)

Multiprocessing function that computes stats for decision tree training

See http://kaldi-asr.org/doc/acc-tree-stats_8cc.html for more details on the Kaldi binary this runs.

Parameters:

directorystr

Directory of training (triphone, speaker-adapted triphone training directories)

align_directorystr

Directory of previous alignment

split_directorystr

Directory of training data split into the number of jobs

ci_phonesstr

Colon-separated list of context-independent phones

num_jobsint

The number of processes to use in calculation

fmllrbool, optional

Whether the current training session is using fMLLR (speaker-adaptation), defaults to False