jose-solorzano / audio-denoiser Goto Github PK
View Code? Open in Web Editor NEWUses machine learning to denoise audio containing speech
License: MIT License
Uses machine learning to denoise audio containing speech
License: MIT License
Hello, I'm having a problem with this package. It seems like it's not using CUDA at all on my machine! I've verified that CUDA is working just fine; in fact, I'm running models on CUDA in the same file as I call this model.
Here's my relevant code. Let me know if you need more details.
import torchaudio
from audio_denoiser.AudioDenoiser import AudioDenoiser
device = "cpu"
if (torch.cuda.is_available()): device = "cuda:0"
# ...
denoiser = AudioDenoiser(device=torch.device(device))
# ...
signal, fs = torchaudio.load(source)
auto_scale = True # Recommended for low-volume input audio
signal = denoiser.process_waveform(waveform=signal, sample_rate=16000, return_cpu_tensor=True, auto_scale=auto_scale)
It seems to work as is, but I'd like to get it to run on CUDA if possible.
Thanks,
Micah
Hello,
I tried both the basic example as well as one with additional parameters, and always getting this error, probably the issue is related to torchaudio library.
(venv) k:\AI\AudioDenoiser>python . K:\AI\AudioDenoiser\sample.wav
k:\AI\AudioDenoiser\__main__.py:9: UserWarning: torchaudio._backend.set_audio_backend has been deprecated. With dispatch
er enabled, this function is no-op. You can remove the function call.
torchaudio.set_audio_backend("soundfile")
Traceback (most recent call last):
File "<frozen runpy>", line 198, in _run_module_as_main
File "<frozen runpy>", line 88, in _run_code
File "k:\AI\AudioDenoiser\__main__.py", line 16, in <module>
denoiser.process_audio_file(in_audio_file, out_audio_file, auto_scale=auto_scale)
File "k:\AI\AudioDenoiser\venv\Lib\site-packages\audio_denoiser\AudioDenoiser.py", line 110, in process_audio_file
waveform, sample_rate = torchaudio.load(in_audio_file)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "k:\AI\AudioDenoiser\venv\Lib\site-packages\torchaudio\_backend\utils.py", line 203, in load
backend = dispatcher(uri, format, backend)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "k:\AI\AudioDenoiser\venv\Lib\site-packages\torchaudio\_backend\utils.py", line 115, in dispatcher
raise RuntimeError(f"Couldn't find appropriate backend to handle uri {uri} and format {format}.")
RuntimeError: Couldn't find appropriate backend to handle uri K:\AI\AudioDenoiser\sample.wav and format None.
(venv) k:\AI\AudioDenoiser>
I see from the error message that some "format" is None, maybe I have to set it explicitly or it's not recognized for some reason.
from sys import argv
import os
from audio_denoiser.AudioDenoiser import AudioDenoiser
import torch
import torchaudio
if __name__ == "__main__":
# Use the "soundfile" audio backend, used in training.
torchaudio.set_audio_backend("soundfile")
# Use a CUDA device for inference if available
device = torch.device('cuda:0') if torch.cuda.is_available() else torch.device('cpu')
denoiser = AudioDenoiser(device=device)
in_audio_file = argv[1]
out_audio_file = f"{os.path.splitext(in_audio_file)[0]}-denoise.wav"
auto_scale = True # Recommended for low-volume input audio
denoiser.process_audio_file(in_audio_file, out_audio_file, auto_scale=auto_scale)
Python: 3.11.1
OS: Windows 10 x64
The WAV file format is a standard: 44.1 kHz, 16 bits, 1 channel, PCM (Little / Signed), Duration: 1 min 43 s
The script is running under venv with audio-denoiser package installed (pip install audio-denoiser
)
Here is the pip list
output:
Package Version
------------------ ----------
audio-denoiser 0.1.1
certifi 2023.11.17
charset-normalizer 3.3.2
colorama 0.4.6
filelock 3.13.1
fsspec 2023.12.2
huggingface-hub 0.20.2
idna 3.6
Jinja2 3.1.2
MarkupSafe 2.1.3
mpmath 1.3.0
networkx 3.2.1
numpy 1.26.3
packaging 23.2
pip 22.3.1
PyYAML 6.0.1
regex 2023.12.25
requests 2.31.0
safetensors 0.4.1
setuptools 65.5.0
sympy 1.12
tokenizers 0.15.0
torch 2.1.2
torchaudio 2.1.2
tqdm 4.66.1
transformers 4.36.2
typing_extensions 4.9.0
urllib3 2.1.0
Appreciate any help.
A declarative, efficient, and flexible JavaScript library for building user interfaces.
๐ Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. ๐๐๐
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google โค๏ธ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.