audio
audio copied to clipboard
pytorch
Data manipulation and transformation for audio signal processing, powered by PyTorch
Reviewed By: dtolnay Differential Revision: D76497877
Add skip condition to vggish_pipeline_test test similar to https://github.com/pytorch/audio/blob/main/test/torchaudio_unittest/backend/dispatcher/ffmpeg/save_test.py#L166 if ffmpeg executable is not found Tested on both rocm and nvidia
This PR fixes the following error: `RuntimeError: torch.linalg.lstsq: only overdetermined systems (input.size(-2) >= input.size(-1)) are allowed on CUDA. Please rebuild with cuSOLVER` Checking the pytorch code https://github.com/ROCm/pytorch/blob/main/aten/src/ATen/native/cuda/linalg/BatchLinearAlgebra.cpp, it calls magma....
Based on PR - https://github.com/pytorch/audio/pull/2485
For rocm, atol is reduced for gradcheck of oscillator_test to pass when using slow mode. Tested on rocm and nvidia
Capture the basic dependencies from https://github.com/pytorch/audio/blob/main/CONTRIBUTING.md#running-test
porbability --> probability follwoing --> following PLEASE NOTE THAT THE TORCHAUDIO REPOSITORY IS NO LONGER ACTIVELY MONITORED. You may not get a response. For open discussions, visit https://discuss.pytorch.org/.
The following tests failed in Python 3.13 when building the conda-forge package (conda-forge/torchaudio-feedstock#30): ``` FAILED test/torchaudio_unittest/backend/dispatcher/soundfile/info_test.py::TestInfo::test_unknown_subtype_warning - assert 2 == 1 + where 2 = len([, ]) ``` This was...
### 🐛 Describe the bug When I try to load a 43-second .wav file, the memory consumption increases, which causes the session to crash. I have about 12GB of RAM....
### 🐛 Describe the bug In [the tutorial for `StreamWriter`](https://docs.pytorch.org/audio/stable/tutorials/streamwriter_basic_tutorial.html#file-like-objects), it is clearly stated that `StreamWriter` works with File-like object that implements `io.RawIOBase.write`. However, when I used `StreamWriter` with the...
