-
Tacotron2: https://arxiv.org/pdf/1712.05884.pdf
WaveRNN: https://github.com/erogol/WaveRNN forked from https://github.com/fatchord/WaveRNN
The idea is to add Tacotron2 as another alternative if i…
-
Hello,
I am trying to train my 20 hours dataset with train_pytorch_tacotron2.v3.yaml. The dataset is correct. It was used many times to synthesize a voice.
Since the 6-th epoch the alignment line …
-
Currently we don't limit the maximum frequency for log-melspectrogram, however, I believe limiting the maximum frequency to 8000kHz or lower (e.g. 7680) alleviates the over-smoothing problem. This wil…
-
I want to reproduce the SOTA TTS result on LJspeech, which based on Transformer.v3 and its MOS is 4.25.
Does "train_pytorch_transformer.v3.yaml" corresponds to the SOTA model configuration? I notice…
-
I observed a interesting behaviour after 138K iters where discriminator dominated the training and generator exploded in both train and validation losses. Do you have any idea why and how to prevent …
-
First of all, thank you for the great implementation of Taco2 with WaveNet.
I'd like to run WaveNet training while I'm still implementing/fixing my text data preprocessing, and I'm planning to mayb…
-
After installing the package nnmnkwii. I tried to run the command
python preprocess.py ljspeech ~/data/LJSpeech-1.1 ./data/ljspeech \
--preset=20180510_mixture_lj_checkpoint_step000320000_ema.js…
-
Hai
Now this github code supporting global and local conditioning features?. I didn't understood for global conditioning use v0.1.1. From current code can i train cmu artic data set by giving g…
-
I'm working on extending TTS to the languages and speakers available in /egs/m_ailabs but synth_wav.sh will only output a default speaker.
I have followed this procedure: executed run.sh through s…
-
In readme i didn't find any steps for VQ-VAE-encoder + WaveNet decoder .Please can any one help me to proceed further.