Found a problem when checking what was to be installed. Also found a temporary solution: mpv's dependency on libpipewire is >= 1.0.4, which means upgrading libpipewire (which causes libspa-0.2-modules to be upgraded) wasn't required. Temporary solution is the older versions of libpipewire and libspa-0.2-modules shown below, remain installed and on hold.

Pkg versions and new Ceres dependency below. Seems Debian decided users need AI.

libpipewire-0.3-0t64 1.4.10-1+b1  >  libpipewire-0.3-0t64 1.6.2-1
libspa-0.2-modules-1.4.10-1+b1   >  libspa-0.2-modules-1.6.2-1

New dependency for libspa-0.2-modules-1.6.2-1 pkg:
libonnxruntime1.23

Purpose and dependencies:
https://packages.debian.org/sid/libonnxruntime1.23

From the Homepage:
https://github.com/microsoft/onnxruntime

ONNX Runtime is a cross-platform inference and training machine-learning accelerator.

ONNX Runtime inference can enable faster customer experiences and lower costs, supporting models from deep learning frameworks such as PyTorch and TensorFlow/Keras as well as classical machine learning libraries such as scikit-learn, LightGBM, XGBoost, etc. ONNX Runtime is compatible with different hardware, drivers, and operating systems, and provides optimal performance by leveraging hardware accelerators where applicable alongside graph optimizations and transforms.

One search result found for onnx at pipewire-org:
https://docs.pipewire.org/devel/page_mo … chain.html
Scroll down to 'ONNX filters'
"There is an optional ONNX filter available..."

Open Neural Network Exchange
https://onnx.ai/

ONNX: Train in Any Framework, Deploy on Any Hardware
Nov 12, 2025
https://www.datacamp.com/tutorial/onnx