FastSpeech2
FastSpeech2 copied to clipboard
Hifi_FS2
all updates for hifi_fs2
Here one error because of the test case set up. We return 3 values from model now to train the discriminator.