Deep GP

The Python Implementation of Deep Gaussian Processes

Branch	Tests	Coverage
main

Currently implemented models are

• Deep GPs
• Variational Auto-encoded Deep GPs (NumPy MLP back-constraint; see CIP-0001)
• Multi-view / MRD Deep GPs (shared latent layer over observation views; see CIP-0002)

Testing

GitHub Actions runs the unit tests (with coverage) on pull requests and pushes to main. The Ubuntu / Python 3.12 job uploads coverage.xml to Codecov.

Requires Python >=3.11 when using GPy 1.15+.

To run the suite locally:

python -m pip install -e ".[test]"
# or, if editable extras are unreliable:
python -m pip install "numpy>=1.7" "scipy>=0.12" "GPy>=1.15" matplotlib h5py tables "coverage>=7.0"
# For psi_chunk_size / psi_threads (CIP-0001 Phase 3) until the next GPy release:
# python -m pip install "git+https://github.com/SheffieldML/GPy.git@main"
python -m pip install -e . --no-deps

python -m coverage run -m unittest discover -s deepgp/testing -p 'test*.py' -v
python -m coverage report

Or with tox:

tox

Coverage is gated on the maintained serial DeepGP path, including the NumPy encoder and MRD (see .coveragerc). MPI parallel inference and prediction-only SVI helpers remain omitted until those paths are removed or revived (CIP-0001 Phases 2–4).

Pre-commit (notebooks)

Commits strip Jupyter outputs and execution counts via nbstripout so notebook diffs stay readable. Install once after cloning::

python -m pip install pre-commit
pre-commit install

Hooks run automatically on git commit. To strip existing notebooks::

pre-commit run nbstripout --all-files

Multi-view / MRD (CIP-0002)

Pass a list of observation matrices and a matching list of output dimensions. The observation-layer kernels must be a list of kernels (one per view)::

m = deepgp.DeepGP(
    [[D1, D2], Q_shared, Q_top],
    Y=[Y1, Y2],
    kernels=[[kern_v1, kern_v2], kern_top],
    num_inducing=40,
    back_constraint=False,
)

With back_constraint=True, encoder_dims[0] is a list of per-view MLP hidden widths (e.g. [[[4], [4]], [3]]); higher indices feed the usual EncoderLayer stack for layers above the MRD leaf. See examples/example_mrd.py.

MPI (deprecated)

Passing mpi_comm to DeepGP emits a FutureWarning (CIP-0001). Use serial models for now; chunked psi-statistics (Phase 3) will replace the single-machine role of MPI. mpi_comm with back_constraint=False raises ValueError (that configuration produced silently wrong results).

An optional GitHub Actions job runs a 2-rank parity check while MPI remains.

Psi chunking (optional)

Large-N models can bound memory and use multiple cores via GPy's chunked RBF / linear psi-statistics (CIP-0001 Phase 3)::

m = deepgp.DeepGP(..., psi_chunk_size=500, psi_threads=4)

Defaults (psi_chunk_size=None, psi_threads=1) preserve previous behaviour. Threading runs inside GPy's pure psi helpers; PyDeepGP itself stays serial at the kernel API.

Examples

• examples/ — short supervised / unsupervised demos (oil, CMU walk/jog).
• examples/damianou2013/ — recreations of the AISTATS 2013 experiments (toy hierarchy, warped regression, USPS depth/ARD, high-five / MRD), including Colab notebooks under examples/damianou2013/notebooks/.