yqzhishen

About audio sample rate and other hparams

1

I noticed that the supported datasets of this repository have different audio sample rate. Is the output sample rate of FastDiff vocoder bound to its training data, or does it...

Will FastDiff work well with singing voice synthesis?

1

Are there any experiments done to evaluate the performance of FastDiff on singing voice datasets?

MIDI SVS mode may produce uneven rhythms

1

Hello and thank you for your great work. However, I tried MIDI SVS of DiffSinger and found that there might be a conceptual mistake in the phoneme duration inference logic,...

The following checks are performed between DiffSinger vocoder and acoustic model: - `sample_rate`, `hop_size`, `num_mel_bins`, `mel_fmin`, `mel_fmax` must be equal - `mel_scale` must be equal and must be `slaney` or...

Some batch edits like DiffSinger autopitch may execute really slowly and block the UI. This PR supports running batch edits asynchronously, with a dialog showing progress of the running batch...

Fixes a tiny mistake in the error message.

[BUG] onnx.onnx_cpp2py_export.checker.ValidationError: Field 'type' of 'value_info' is required but missing.

2

**Describe the bug** ONNX Simplifier fails to simplify models with optional types. The following minimum demo: ```python from typing import Optional import onnx import onnxsim import torch class MyModule(torch.nn.Module): def...

The former iSTFT code has a bug which will produce wrong results when batch_size > 1: ```python coffidx = th.where(coff > 1e-8) outputs[coffidx] = outputs[coffidx]/(coff[coffidx]) ``` When you use `torch.where(condition)`,...