True. But it really becomes a pain to observe the changes for a single speaker TTS training runs.

→ View original post on X — @reach_vb