-
Hi, I'm currently updating my torch codebase from using librosa to torchaudio for transforms, to take advantage of the (much) faster stft torch implementation on the GPU. However, running into several…
-
I've been writing a pytorch frontend, and I'm currently porting mfsc featurization.
I just noticed SpeechUtils scales the float samples to integer range:
https://github.com/facebookresearch/wav2le…
-
## 🐛 Bug
## To Reproduce
Steps to reproduce the behavior:
1. take input tensor of size `torch.Size([64, 276978])`
2. forward pass with
```
torchaudio.transforms.MFCC(sample_rate=16000…
-
Lazy initialization refers to [the declaration here](https://github.com/pytorch/audio/blob/a424509dda5b57c932fa8b5b780de93e60ed7ee2/torchaudio/transforms.py#L135) and [the initialization here](https:/…
zh217 updated
5 years ago
-
Hi,
We are doing STT / TTS for the Russian [language](https://github.com/snakers4/open_stt).
We mostly used STFT due to our ignorance in DSP and our understanding that MFCC filters used by every…
-
Not sure what's going on here, but I've submitted a code example:
import torch
import torchaudio
class Example(torch.nn.Module):
def __init__(self, time_seconds: int = 1, sr: int = 1600…
-
can we synthesize directly from melspectrogram without using encoder
-
Hello, @rafaelvalle Really nice work.
I would like to combine this work with global style tokens.. please let me know
-
in nikl_m.py
mel_spectrogram = audio.melspectrogram(wav).astype(np.float32)
https://r9y9.github.io/nnmnkwii/stable/references/generated/nnmnkwii.preprocessing.preemphasis.html
Traceback…
-
Finally I got the _python -m multiproc train.py etc..._ to work. Simple question: How do I now synthesize audio form a specific checkpoint?
In github.com/keithito/tacotron it was pretty simple: _p…