qalgora-QX · Solvers Library

The Solvers library packages ready-made hybrid algorithms — VQE, ADAPT-VQE, QAOA, and gradient routines — so you call a high-level solver instead of wiring the loop by hand.

Installation

# python3 -m pip install qalgora-solvers   # planned — not yet installable

Ground-state energy with ADAPT-VQE

import qalgora
import qalgora_solvers as solvers
from qalgora import spin

# molecular Hamiltonian (e.g. from a chemistry frontend)
hamiltonian = spin.z(0) + spin.z(1) - 0.5 * spin.x(0) * spin.x(1)

# operator pool the solver draws from to grow the ansatz
pool = solvers.get_operator_pool("spin_complement", num_qubits=2)

energy, params, ops = solvers.adapt_vqe(
    hamiltonian, pool,
    optimizer=solvers.optimizers.LBFGS(),
    gradient="parameter_shift",
)
print("ground-state energy:", energy)

Fixed-ansatz VQE

import qalgora_solvers as solvers

# molecule -> spin Hamiltonian + a UCCSD ansatz, in one call
h, data = solvers.create_molecule(geometry, basis="sto-3g")
energy, params = solvers.vqe(h, solvers.uccsd, data.num_qubits,
                             optimizer=solvers.optimizers.LBFGS())

Generative eigensolver (GQE)

Instead of optimising parameters of a fixed ansatz, the Generative Quantum Eigensolver trains a transformer to generate low-energy circuits from an operator pool — a gradient-free, sampling-based alternative to VQE for hard energy landscapes.

import qalgora_solvers as solvers

energy, circuit = solvers.gqe(
    hamiltonian, pool,
    model=solvers.gqe.Transformer(layers=4),
    iterations=200,
)

Operator pools

ADAPT-VQE grows its ansatz from a pool. Choose a pool that matches your problem's symmetry.

What's included