Fix condition_dp_on_speaker

This commit is contained in:
Eren G??lge 2022-06-18 14:03:46 +02:00
parent 18dbf22195
commit a969a11c04
1 changed files with 8 additions and 0 deletions

View File

@ -504,6 +504,12 @@ class VitsArgs(Coqpit):
to the `config.audio.sample_rate`. If it is False you will need to add extra
`upsample_rates_decoder` to match the shape. Defaults to True.
reinit_DP (bool):
Reinitialize the duration predictor weights at the beginning of a training. Defaults to False.
reinit_text_encoder (bool):
Reinitialize the text encoder weights at the beginning of a training. Defaults to False.
"""
num_chars: int = 100
@ -876,6 +882,8 @@ class Vits(BaseTTS):
attn = maximum_path(logp, attn_mask.squeeze(1)).unsqueeze(1).detach() # [b, 1, t, t']
# duration predictor
if not self.args.condition_dp_on_speaker:
g = None
attn_durations = attn.sum(3)
if self.args.use_sdp:
loss_duration = self.duration_predictor(