two pitch?

[a897456]

the first pitch in the sample() as follow:

naturalspeech2-pytorch/naturalspeech2_pytorch/naturalspeech2_pytorch.py

Lines 1478 to 1479 in 659bec7

	duration, pitch = self.duration_pitch(phoneme_enc, prompt_enc)
	pitch = rearrange(pitch, 'b n -> b 1 n')

the second pitch in the forward() of Naturalspeech2 as follow:

naturalspeech2-pytorch/naturalspeech2_pytorch/naturalspeech2_pytorch.py

Lines 1543 to 1556 in 659bec7

	if not exists(pitch):
	assert exists(audio) and audio.ndim == 2
	assert exists(self.target_sample_hz)
	if self.calc_pitch_with_pyworld:
	pitch = compute_pitch_pyworld(
	audio,
	sample_rate = self.target_sample_hz,
	hop_length = self.mel_hop_length
	)
	else:
	pitch = compute_pitch_pytorch(audio, self.target_sample_hz)
	pitch = rearrange(pitch, 'b n -> b 1 n')

1. Personally, I think the first pitch is from the prompt, and the second pitch is from the training data, right?
2. Personally, I think the prompt is a small part of the training data, such as the training data is10s, from which prompt takes 2s, right?
3. Because the input format of the prompt and the training data is the same, why are the calculation methods of pitch different?

[lexkoro]

One is the ground truth pitch and the other one is the predicted

[a897456]

One is the ground truth pitch and the other one is the predicted

Thank you for your reply. @lexkoro
By the way, have you completed the conditional training?
And can you share how to generate the prompt and the text just like LJSpeech dataset.

[lexkoro]

I don't think the repository is usable yet.