LEE CHOONGHO

Hello, I'm trying to implement Differentiable Duration Modeling(DDM) module introduced in [Differentiable Duration Modeling for End-to-End Text-to-Speech](https://arxiv.org/abs/2203.11049). I opened this issue to get advice on implementation DDM. My Implementation of...

Training issue

5

Thanks for sharing the nice model implementation.  When I start training, the following warning appears, do you also get the same message? I think it's a fairseq installation problem....

Question about noise scheduling process.

1

Hello I'm trying to implement noise scheduling process refer to BDDM's implementation [BDDM/sampler.py](https://github.com/tencent-ailab/bddm/blob/main/bddm/sampler/sampler.py) And I have some question for noise scheduling process for FastDiff-TTS. 1. In the Fastdiff paper, the...

Stochastic duration prediction failed for fastspeech2

4

I applied the stochastic duration predictor to the fastspeech2 model. Duration loss is falling smoothly (1.2 to 0.2)  But, in inference, the duration predictor does not work at all....

Is there any research about appropriate size of dataset for Fastspeech2?

2

I'm training Fastspeech2 with multi-lingual TTS Dataset like below. - Number of data : 300000 English(~44000) + Chinese(~80000) + Spanish(~30000) + Japanese(~7000) + Korean(~130000) : Total ~300000 - Number of...

**Describe the bug** I operate pytorch training server with 2 celery worker, redis broker, and fastapi server. And when I give several tasks(3~4) to workers, tasks are well reserved and...

32kHz Vocos Multi Speaker Model Training Log

13

Training Loss, Generated Outputs. I hope this will be a reference for model training. https://api.wandb.ai/links/xi-speech-team/k0kdfwch

Question about pitch loss

12

Hello, thank you for sharing your work. In the paper, pitch loss is calculated by formula like below `pitch_loss = huber_norm(log2(pred_shifted_f0) + 0.5 * d, log2(pred_f0), delta=???)` But I can't...