Note:

KALDI-ASR recipe to clone whenever I need! run.sh to build and decode various acoustic models - monophone, triphone, DNN-HMM, LDA-MLLT, LSTM, BLSTM.

unified_run.sh covers all - data-preparation, features-extraction, model-training & decoding, saving-results & decoding

Run transliterator using $sudo perl path-to/transliterator.pl path-to/english_tam_map path-to-the-file-that-has/Tamil word path-to/out_word

Name		Name	Last commit message	Last commit date
Latest commit History 27 Commits
conf		conf
local		local
steps		steps
utils		utils
Create_ngram_LM.sh		Create_ngram_LM.sh
KD.py		KD.py
NM.py		NM.py
README.md		README.md
Tamizh-map		Tamizh-map
best_score.sh		best_score.sh
cmd.sh		cmd.sh
data_prep1.sh		data_prep1.sh
english_tam_map		english_tam_map
lexicon.txt		lexicon.txt
mfcc.ipynb		mfcc.ipynb
old-run.sh		old-run.sh
path.sh		path.sh
phone_list		phone_list
phoneset_all		phoneset_all
phoneset_mei		phoneset_mei
phoneset_uyir		phoneset_uyir
phoneset_uyirmei		phoneset_uyirmei
prep_lexicon.py		prep_lexicon.py
proper_split.py		proper_split.py
reduce_lexicon.py		reduce_lexicon.py
results.txt		results.txt
score.sh		score.sh
syllable_from_char		syllable_from_char
test10.wav		test10.wav
testing_dysarthric_asr.sh		testing_dysarthric_asr.sh
text_prep.py		text_prep.py
transliterator.pl		transliterator.pl
unified_run.sh		unified_run.sh
unique_Tamizh_words		unique_Tamizh_words
unique_Tamizh_words (copy)		unique_Tamizh_words (copy)

Provide feedback