How flagos-ai/flagquantum works
FlagQuantum is a PyTorch-first framework for differentiable quantum computing. Its principal workflow is to define a circuit and requested observable, train circuit parameters, then evaluate the trained circuit through a selected local or remote target and inspect the result. The repository also includes compiler and execution-planning layers, several simulation representations, hardware/provider integrations, QPU digital twins, and quantum-error-correction research tools. Some execution and integration internals were not sampled, so those connections are shown only where documented or are marked unverified.
Main entry points: execution_plan.py Execution planning · _api.py Public run API · circuit.py Circuit API · local.py Statevector simulation
Components 20
Circuit authoring
- Circuit API
flagquantum/circuit.py - Circuit IR
flagquantum/core/ir.py - Observable requests
flagquantum/observables/__init__.py - PyTorch training
flagquantum/runtime/training.py
Execution runtime
- Public run API
flagquantum/_api.py - Execution planning
flagquantum/runtime/execution_plan.py - Circuit compiler
flagquantum/compiler/pipeline.py - Execution result
flagquantum/runtime/result.py
Execution targets
- Statevector simulation
flagquantum/simulation/statevector/local.py - Tensor-network simulation
flagquantum/simulation/tensor_network/models.py - MPS simulation
flagquantum/simulation/mps/entrypoints.py - Distributed training
flagquantum/runtime/executors/mps/training_engine.py - Compute runtimes
flagquantum/compute/registry.py - Jiuding workspace
flagquantum/remote/compute/jiuding.py - Remote QPU
flagquantum/remote/qpu/execution.py - SDK adapters
flagquantum/ecosystem/registry.py
Research capabilities
- Error correction
flagquantum/qec/circuit.py - QPU digital twins
flagquantum/twin/candidate.py - Quantum algorithms
flagquantum/algorithms
Other components
- Quantum developer
Connections 21
- Quantum developer → Circuit API: defines circuit
README.md - Quantum developer → Observable requests: chooses outputs
README.md - Circuit API → Circuit IR: records operations
flagquantum/circuit.py - Circuit API → Statevector simulation: initializes state
flagquantum/circuit.py - Quantum developer → PyTorch training: trains model
README.md - Quantum developer → Public run API: runs circuit
README.md - Public run API → Execution result: returns result
flagquantum/_api.py - Public run API → Execution planning: plans execution
flagquantum/_api.py - Public run API → Circuit compiler: compiles circuit
- Execution planning → Statevector simulation: selects simulator
- Execution planning → Tensor-network simulation: selects simulator
- Execution planning → MPS simulation: selects simulator
- Execution planning → Jiuding workspace: selects target
- Execution planning → Remote QPU: selects target
- Compute runtimes → Distributed training: provides runtime
- Quantum developer → Jiuding workspace: requests simulation
README.md - Quantum developer → Remote QPU: requests hardware run
README.md - Quantum developer → Error correction: explores QEC
README.md - Quantum developer → QPU digital twins: compares device data
README.md - Quantum developer → Quantum algorithms: uses algorithms
- Quantum developer → SDK adapters: uses SDK adapters