Keras implementation of Tamas Gabor Csapo, Csaba Zainko, Laszlo Toth, Gabor Gosztolya, Alexandra Marko, ,,Ultrasound-based Articulatory-to-Acosutic Mapping with WaveGlow Speech Synthesis'', accepted at Interspeech 2020, arXiv:2008.03152
For WaveGlow synthesis, download model from https://drive.google.com/file/d/1cjKPHbtAMh_4HTHmuIGNkbOkPBD9qwhj/view?usp=sharing
Pretrained models and synthesized samples (May 7, 2020) at https://simonyi-my.sharepoint.com/:f:/g/personal/csapszi_sch_bme_hu/EjBcbCcVUmFCkhh_NiVOZikBB8CsyATBVQA9aaGK-jhh3A?e=vlc50W
Requirements
- Python 3.6.9
- keras 2.2.4
- tensorflow-gpu 1.13.1
- pytorch 1.2.0
- ...etc... see requirements.txt