Is there a free, open-source or even non-HTK licensed version of this toolkit? Is there another way to adapt models for Julius?
>Is there a free, open-source or even non-HTK licensed version of this toolkit?
Not that I know of...
You might try creating/adapting your acoustic models using SphinxTrain, and then converting the Sphinx acoustic model to htk acoustic model...
Or, just use SphinxTrain for training acoustic models, and one of the Sphinx group recognition engines...