Skip to content

temporary fix on the torchaudio IO #80

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
YunchaoYang wants to merge 2 commits into
base: main
Choose a base branch
from
Open

temporary fix on the torchaudio IO #80

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
7177372
temporary fix on the torchaudio IO
YunchaoYang Feb 2, 2026
b6919d2
move scipy dependency to dev only
YunchaoYang Feb 4, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
5 changes: 3 additions & 2 deletions pyproject.toml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -51,7 +51,8 @@ dependencies = [
"mypy>=0.782",
"pylint>=2.8.0",
"flake8",
"types-tqdm"
"types-tqdm",
"scipy>=1.7",
]
cpu = [
"torch==2.6.0+cpu",
Expand Down Expand Up @@ -135,4 +136,4 @@ explicit = true
[[tool.uv.index]]
name = "pytorch-cpu"
url = "https://download.pytorch.org/whl/cpu"
explicit = true
explicit = true
8 changes: 6 additions & 2 deletions tests/integration_tests/test_sonar_speech_pipeline_models.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -6,8 +6,9 @@

from pathlib import Path

import numpy as np
import torch
import torchaudio # type: ignore
from scipy.io import wavfile
from torch.testing import assert_close # type: ignore

from sonar.inference_pipelines.speech import (
Expand All @@ -21,8 +22,11 @@
str(DATA_DIR.joinpath("audio_files/audio_2.wav")),
]

WAV, sr = torchaudio.load(AUDIO_Paths[0])
# Use scipy for WAV file reading (torchaudio I/O removed in 2.9+)
sr, audio_numpy = wavfile.read(AUDIO_Paths[0])
assert sr == 16000, "Sample rate should be 16kHz"
# Convert int16 to float32 and add channel dimension
WAV = torch.from_numpy(audio_numpy.astype(np.float32) / 32767.0).unsqueeze(0)


def test_speech_to_embedding_model_pipeline():
Expand Down
8 changes: 5 additions & 3 deletions tests/unit_tests/test_sonar_speech.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -6,9 +6,10 @@

from pathlib import Path

import numpy as np
import pytest
import torch
import torchaudio # type: ignore[import]
from scipy.io import wavfile

from sonar.inference_pipelines.speech import SpeechToEmbeddingModelPipeline

Expand Down Expand Up @@ -39,9 +40,10 @@ def test_speech_embedding_with_waveform_input(encoder):

# Parsing audio within sonar does not support fp16 audio decoding yet
def test_speech_embedding_pipeline_with_audio_files(tmp_path: Path, encoder):
print(torchaudio.list_audio_backends())
fake_audio = torch.rand(1, 175920)
audio_file = tmp_path / "audio.wav"
torchaudio.save(audio_file, fake_audio, 16000)
# Use scipy for WAV file writing (torchaudio I/O removed in 2.9+)
audio_numpy = fake_audio.squeeze().numpy()
wavfile.write(str(audio_file), 16000, (audio_numpy * 32767).astype(np.int16))
embedding = encoder.predict([str(audio_file.resolve())])
assert embedding.shape == torch.Size([1, 1024])
Loading