This is a kaldi based recipie for Malayalam speech recognition. You need a working Kaldi directory to run this script.
Details on how to run this script and the working is described here.
To install Kaldi, see the documentation here
The source code of /asr_malayalam
has to be placed in the /egs
directory of Kaldi installation directory.
./run.sh ./inputdirectory
input directory has the following structure:
├── language -> symlink/to/language/model/source ( lm_train.txt, lexicon.txt)
├── test
│ └── corpus1
│ ├── audio -> symlink/to/audio/files_directory (utt1.wav, utt2.wav)
│ └── metadata.tsv -> symlink/to/audio/files/metadata (metadata.tsv)
└── train
├── corpus2
| ├── audio -> symlink/to/audio/files_directory (utt1.wav, utt2.wav)
| └── metadata.tsv -> symlink/to/audio/files/metadata (metadata.tsv)
└── corpus3
├── audio -> symlink/to/audio/files_directory (utt1.wav, utt2.wav)
└── metadata.tsv -> symlink/to/audio/files/metadata (metadata.tsv)
metadata.tsv is a tab separated values of utterence_is, speaker_id, file_name in audio folder, transcript in Malayalam script.