blueqat documentation

Note

日本語版ドキュメントは こちら (Japanese documentation is also available).

blueqat is an open-source Python SDK for quantum computing, built natively on PyTorch. Circuits run on a differentiable statevector / tensor-network simulator, so gradients of quantum programs (for VQE, QAOA, pulse optimization, …) come for free through autograd.

from blueqat import Circuit

# A Bell pair, sampled 100 times
Circuit(2).h[0].cx[0, 1].m[:].run(shots=100)
# => Counter({'00': 52, '11': 48})

Highlights

  • Two execution modes behind one API: dense statevector and tensornet (tensor-network contraction, the default) for large circuits.

  • Differentiable end to end: gate parameters can be torch.Tensor values with requires_grad=True.

  • Exchange-only spin qubits (blueqat.eo): encode logical qubits in 3 spins and compile circuits to Heisenberg exchange pulses, including differentiable pulse synthesis and hardware-facing pulse schedules.

  • Interop: OpenQASM 2.0 input/output, versioned JSON circuit serialization, circuit drawing.

  • Cloud groundwork (blueqat.cloud): API-key management and a backend='cloud' submission path.

User guide

API reference

日本語 (Japanese)

Indices and tables