This project is being developed in Python 3.6, but is tested on Python 2.7 and Python 3.6. I've tried my best to make it compatible with non-Python 3.6 versions, but I make no guarantees. In ...

Abstract: LVCSR based audio indexing approaches are preferred as they allow search, navigation, browsing and structuring of audio/video documents based on their content. A major challenge with LVCSR ...

Abstract: The diachronic nature of broadcast news data leads to the problem of out-of-vocabulary (OOV) words in large vocabulary continuous speech recognition (LVCSR) systems. Analysis of OOV words ...

In this paper, we evaluate the effect of OOV detection and recovery for a low-resource language on WER. We use a small seed corpus of continuous speech and improve the vocabulary by incorporating the ...

More up-to-date version of wrapper is here. You should use that. This has code which wraps OpenFST, and functions for modifying graphs (insert, replace_single, add_boost). To compile you will need to ...