Index of /opt/mod9-asr/models/asr/tuda/swc_mailabs_cv_voc683k_smaller_fst/graph

 Name                                                              Size  Description
 Parent Directory                                                    -   Decoding graph (composing language model) with lexicon and optional phonetic inventory.
 HCLG.fst.gz                                                       653M  Compressed serialization of OpenFST data structure (ASR Engine will decompress at runtime).
 words.txt                                                          12M  Mapping of words to be transcribed from graph-specific output labels; this lexicon/vocabulary could be manually edited.
 phones.txt                                                        4.3K  Mapping of phone(me)s to be transduced as graph-specific input labels; this is needed for model customization.
 phones/                                                             -   Mapping of word-position-dependent phones to enable word-alignment; this is needed for word-level timestamps.