GitLab

trunk/egs/hkust: add alternative chinese character scoring scripts; if the wordlist (words in the lexicon) used by the decoder is a subset of the words in 195k_chinese_word2char_map (1st column of 195k_chinese_word2char_map contains chinese words, 2nd to end column fields are corresponding seperated chinese characters), scoring with score_basic_ext.sh will give pure chinese character error rate

git-svn-id: https://svn.code.sf.net/p/kaldi/code/trunk@2748 5e6a8d80-dfce-4ca6-a32a-6e07a63d50c8