Consistency of input frame Leno and generated waveform length

[rita-zeng]

I wonder how to ensure the consistency of input frame length and output waveform length ? When I use GRID datasets to train and test and set the hyper parameters as follow:
T = 40
overlap = 10
mel_step_size = 160
mel_overlap = 40
img_size = 96
fps = 25,
Test results shows that the ground truth is 3 seconds while the generated waveforms are 7 seconds. How can I solve this problem? Looking forward to your reply!