Welcome to FastSpeech2 Discussions!

[rishikksh20]

👋 Welcome!

We’re using Discussions as a place to connect with other members of our community. We hope that you:

• Ask questions you’re wondering about.
• Share ideas.
• Engage with other community members.
• Welcome others and are open-minded. Remember that this is a community we
  build together 💪.

To get started, comment below with an introduction of yourself and tell us about what you do with this community.

[luudinh93]

i am training your model but getting error at step hs=hs+self.pitch_embeded because their dimensionality is different. Please help me fix this error

[rishikksh20]

length of hidden state (1261) and length of the pitch(1119) is different, both should be same either 1261 or 1119. Please check your mel files

[luudinh93]

I have fixed the above error but there is another error

[rishikksh20]

check your duration for respective phonemes I think some duration in your pre-processed file is negative

[luudinh93]

I guess I created the filelist file the wrong way. Please give me detailed instructions for this step

Example:
tgs, tss = textgrid.TextGrid.fromFile(path_to_textgrid)

tgs = IntervalTier(words, [Interval(0.0, 0.46, None), Interval(0.46, 0.64, chú), Interval(0.64, 1.09, thích), Interval(1.09, 1.11456, None)])

tts = IntervalTier(phones, [Interval(0.0, 0.46, None), Interval(0.46, 0.55, ch), Interval(0.55, 0.64, u1_T3), Interval(0.64, 0.73, th), Interval(0.73, 0.81, i_T3), Interval(0.81, 1.09, ch), Interval(1.09, 1.11456, None)])

File that I created:
metadata text: audio_001.wav|chú thích.
<unk> chú thích <unk>| -2 0 46 55 64 73 81 109|2 46 9 9 9 8 28|pau ch u1_T3 th i_T3 ch pau|audio_001.wav