How we can slow down the output audio?

[ayush431]

No description provided.

[raccoonML]

A more advanced solution is to save the attention layer alignments from inference, stretch them by the desired slowdown amount, then run the decoder loop again replacing the attention network output with the stretched alignments. This has the effect of forcing the decoder to use more frames when synthesizing each letter of every word, slowing down the output. With similar techniques, it's also possible to slow down selected words, or speed up.