Hi everyone! As you might recall, we’re hard at work on the upcoming Mimic 2 voices. We’ve been busy sampling a new voice, courtesy of my colleague @kusal-mycroft. My colleague @Mn0491 has been working super-hard tuning hyper-parameters in Tensorflow to help train the machine learning network to produce a clean, easy to listen to, and life-like voice.
While progress has been steady, one of the challenges we’ve encountered has been the time it takes to train each epoch.
Is anyone aware of any voice post-production software that works on lower-end hardware? That way we can better train the voice - and have new voices ready for you quicker