| *** using default audio.py *** | |
| 60 epoches | |
| 4589 steps per epoch | |
| 295k steps | |
| 0.8 step per second | |
| loss = 0.21 | |
| ################################ | |
| *** using new audio.py (to be compatabel with hifigan) *** | |
| * epochs | |
| 2876 steps per epoch | |
| 214 steps | |
| 0.65 step per second | |
| loss = 0.62 | |
| ############################### | |
| *** using default audio.py and resegmented wavs *** | |
| * epochs | |
| 2876 steps per epoch | |
| 320k steps | |
| 0.63 step per second | |
| loss = 0.28 | |
| ############################## | |
| *** using default audio.py with changing in some parameters and resegmented wavs *** | |
| * epochs | |
| 2868 steps per epoch | |
| 320k steps | |
| 0.64 step per second | |
| loss = 0.28 | |