sci-ml/julius
Differentiable digital signal processing for PyTorch
Screenshots
View
Download
BrowseChangeLog
commit 67ad6cbce8b715f55cb18428c48e095b909bcffa
Author: Ivan S. Titov <iohann.s.titov@gmail.com>
Date: Sat May 16 09:42:00 2026 +0200
sci-ml/julius: new package, add 0.2.7
PyTorch DSP utilities: differentiable resampling, FFT convolutions,
filterbanks. Imported as a transitive runtime dep of
sci-ml/torch-audiomentations on the pyannote.audio path.
Upstream is 2022-vintage and dormant on PyPI; the 0.2.7 release
fixed ONNX export and remains the latest. SINGLE_IMPL inherited
from sci-ml/pytorch; tests restricted (need resampy + onnxruntime
+ audio fixtures we don't carry).
Author: Ivan S. Titov <iohann.s.titov@gmail.com>
Date: Sat May 16 09:42:00 2026 +0200
sci-ml/julius: new package, add 0.2.7
PyTorch DSP utilities: differentiable resampling, FFT convolutions,
filterbanks. Imported as a transitive runtime dep of
sci-ml/torch-audiomentations on the pyannote.audio path.
Upstream is 2022-vintage and dormant on PyPI; the 0.2.7 release
fixed ONNX export and remains the latest. SINGLE_IMPL inherited
from sci-ml/pytorch; tests restricted (need resampy + onnxruntime
+ audio fixtures we don't carry).