* Fix import/export of model * update speech sample * fix code-style Co-authored-by: jiwaszki <jan.iwaszkiewicz@intel.com>