Qumode-based Variational Quantum Deflation (QumVQD): a framework for computing electronic and vibrational excited-state energies on bosonic / qumode quantum hardware models. QumVQD maps molecular Hamiltonians onto one or more qumodes and optimizes a SNAP + displacement + beamsplitter ansatz in TensorFlow, recovering ground and excited states through an overlap-penalty based cost function.
helpers/QumVQD_Helpers.py Core ansatz, VQD loss, optimizer, Hamiltonians
third_party/vibrational_dynamics/ Vibrational Hamiltonian fragmentation code
notebooks/
Electronic_QumVQD.ipynb Electronic excited states for H2 / H4
Vibrational_VQD.ipynb Vibrational excited states
Manuscript_Plotting.ipynb Figures for the manuscript
Plots/ Stores plots produced by
data/ Data required to run notebooks
pyproject.toml Package metadata (installable as `QumVQD`)
NOTICE.md Third-party attribution
.gitattributes Handles Git LFS configuration
.gitignore Designates which files for Git to ignore
LICENSE.md MIT License
THIRD_PARTY_LICENSES.md Attribution for third-party code
Requires Python ≥ 3.9.
Warning: Windows users must use Windows Subsystem for Linux due to dependency on PySCF. Moreover, this repository uses Git LFS for the .hdf5 data files. Install it first (git lfs install), then clone.
git clone https://github.com/sdonglab/QumVQD.git
cd QumVQD
pip install -e .This installs the helpers and third_party packages along with the
dependencies declared in pyproject.toml (QuTiP, TensorFlow, TensorFlow
Probability, OpenFermion, openfermionpyscf, scikit-learn, Jupyter).
Open one of the notebooks under notebooks/ and run top-to-bottom. Key run
parameters at the top of Electronic_QumVQD.ipynb:
| Parameter | Meaning |
|---|---|
molecule |
"H2" or "H4" |
hhdis |
H–H bond length in Å |
num_states |
Number of states to compute (ground + excited) |
ndepth |
Number of ansatz layers |
nfocks |
List of Fock-space dimensions per qumode, e.g. [70] or [16,16] |
beta_factor |
Weight of the orthogonality penalty in the VQD cost |
niter |
Maximum optimizer iterations per state |
threshold |
Convergence threshold on the loss |
This repository includes code adapted from:
Both are used under the MIT License. See NOTICE.md, the
third_party/ directory, and THIRD_PARTY_LICENSES.md
for full attribution and license text.
MIT — see LICENSE.
Marlon F. Jost, Sijia S. Dong. Excited-State Quantum Chemistry on Qumode-Based Processors via Variational Quantum Deflation. arXiv. 2026. arXiv:2604.13457
@misc{jost2026excitedstatequantumchemistryqumodebased,
title={Excited-State Quantum Chemistry on Qumode-Based Processors via Variational Quantum Deflation},
author={Marlon F. Jost and Sijia S. Dong},
year={2026},
eprint={2604.13457},
archivePrefix={arXiv},
primaryClass={quant-ph},
url={https://arxiv.org/abs/2604.13457},
}
Jost, M. and Dong, S. (2026) “QumVQD: Qumode-Based Variational Quantum Deflation Framework”. Zenodo. DOI:10.5281/zenodo.20479296