audio

audio copied to clipboard

[Codemod][Remove unused Python imports] pytorch/audio 3/6

3

Differential Revision: D53400249

azad-meta

torchaudio.functional.convolve with torch.complex32 and interleaved fake complex torch.bfloat16

1

### 🚀 The feature I have a use case for convolving two tensors with dtypes torch.float16 or torch.bfloat16 containing interleaved complex data. ### Motivation, pitch I have a couple DSP...

pfeatherstone

`torchaudio.functional.rnnt_loss` crashes for `logits` with >2**31 elements

2

### 🐛 Describe the bug Repro: ```python import subprocess import torch import torchaudio b, t, u, v = 33, 256, 64, 4096 print( torchaudio.functional.rnnt_loss( torch.zeros((b, t, u, v), dtype=torch.float16, device='cuda'),...

gpuplz

torchaudio.io.StreamReader crashes on repeated seek / next

1

### 🐛 Describe the bug If I repeatedly run the following script, it will fail before reaching `"End"` most of the time. About 75% of the time I run the...

MichaelCurrie

specify fmin and fmax for Spectrogram

2

### 🚀 The feature specify `fmin` and `fmax` for Spectrogram like `MelSpectrogram`. ### Motivation, pitch We can specify `fmin` and `fmax` for `MelSpectrogram`, but we cannot for `Spectrogram`. If we...

bilzard

[IGNORE] debugging

1

Please ignore, I'm just trying to debug a PR in pytorch core https://github.com/pytorch/pytorch/pull/107131 that made some torchaudio tests fail https://github.com/pytorch/pytorch/issues/107531

NicolasHug

librispeech_conformer_rnnt example gives error: "AttributeError: 'tuple' object has no attribute 'targets'"

1

### 🐛 Describe the bug Hi I am trying to run locally the librispeech_conformer_rnnt ASR example given [here](https://github.com/pytorch/audio/tree/main/examples/asr/librispeech_conformer_rnnt). I installed the nightly versions of PyTorch and TorchAudio and the mentioned...

kikofmas

Resampling at arbitrary time steps

5

### 🚀 The feature Currently, `torchaudio.functional.resample` can only resample at regular time points and the period is determined by `orig_freq` and `new_freq`. Is it possible to resample at arbitrary time...

pfeatherstone

`torchaudio.info` returns incorrect result for `num_frames` when input is a video

1

### 🐛 Describe the bug If the input is a video, `torchaudio.info().num_frames` returns the incorrect result. For example: ```python import torchaudio from subprocess import check_call url = "https://download.pytorch.org/torchaudio/tutorial-assets/stream-api/NASAs_Most_Scientifically_Complex_Space_Observatory_Requires_Precision-MP4_small.mp4" check_call(["wget", url,...

lematt1991

### 📚 The doc issue https://pytorch.org/audio/stable/generated/torchaudio.transforms.Vol.html The current doc for `torchaudio.transforms.Vol` just says "Adjust volume of waveform." but it also clamps the output using `torch.clamp(waveform, -1, 1)` which can be...

CookiePPP