技能 pennylane

🔮

pennylane

Name: pennylane
Author: davila7

安全 ⚙️ 外部命令🔑 環境變數🌐 網路存取📁 檔案系統存取

Build quantum circuits with PennyLane

也可從以下取得: K-Dense-AI

Quantum computing tasks require specialized knowledge of circuit design and optimization. PennyLane provides a unified interface for building, training, and executing quantum circuits with automatic differentiation across simulators and real hardware.

支援: Claude Codex Code(CC)

📊 71 充足

下載技能 ZIP

在 Claude 中上傳

前往設定 → 功能 → 技能 → 上傳技能

開啟並開始使用

測試它

正在使用「pennylane」。 Create a simple variational quantum classifier that classifies two-dimensional data points

預期結果:

Define a 2-qubit device with default.qubit backend
Build a QNode with angle encoding for input features
Create variational layers with trainable weights
Use expectation value of PauliZ as classification output
Train with GradientDescentOptimizer minimizing mean squared error
Evaluate accuracy on training and test datasets

正在使用「pennylane」。 Show me how to run VQE for H2 molecule

預期結果:

Import qchem module and define atomic coordinates for H2
Build molecular Hamiltonian using qchem.molecular_hamiltonian
Initialize Hartree-Fock state for the molecule
Create VQE ansatz with UCCSD (unitary coupled cluster) template
Run optimization loop using AdamOptimizer
Extract ground state energy in Hartree units

正在使用「pennylane」。 How do I switch between simulator and real quantum hardware

預期結果:

Define your quantum circuit as a function accepting a device parameter
Create simulator device: qml.device('default.qubit', wires=4)
Create hardware device: qml.device('qiskit.ibmq', wires=4, backend='ibmq_manila')
Pass the device to your QNode decorator
Circuit code remains identical - only the execution backend changes
Use parameter-shift rule for gradients on hardware (backprop only works on simulators)

安全審計

安全

v5 • 1/17/2026

Pure documentation skill containing educational Python code examples for quantum computing workflows. All 419 static findings are false positives: code examples in markdown files are documentation, not executable code. The skill has no scripts, no network calls, no file system access, and no real credentials. The api_key placeholders are documentation examples, not actual secrets. Quantum computing algorithms flagged as 'weak crypto' are legitimate scientific computing patterns.

已掃描檔案

4,186

分析行數

發現項

審計總數

風險因素

⚙️ 外部命令 (354)

🔑 環境變數 (2)

references/devices_backends.md:255 references/devices_backends.md:255

🌐 網路存取 (10)

references/getting_started.md:224 references/getting_started.md:225 references/getting_started.md:226 references/getting_started.md:227 skill-report.json:6 SKILL.md:216 SKILL.md:217 SKILL.md:218 SKILL.md:219 SKILL.md:220

📁 檔案系統存取 (2)

references/optimization.md:414 references/optimization.md:417

審計者: claude 查看審計歷史 →

品質評分

架構

100

可維護性

內容

社群

100

安全

規範符合性

你能建構什麼

Quantum chemistry simulations

Calculate molecular ground state energies using VQE with UCCSD ansatz for drug discovery research

Hybrid quantum-classical models

Build variational classifiers combining quantum circuits with PyTorch or TensorFlow layers

Optimization algorithms

Implement QAOA for combinatorial optimization problems like MaxCut on graphs

試試這些提示

Basic circuit

Show me how to create a 2-qubit quantum circuit in PennyLane that applies RX and RY rotations then measures PauliZ on qubit 0

VQE calculation

Write a PennyLane script to compute the ground state energy of H2 molecule using VQE with Adam optimizer

Hybrid classifier

Create a hybrid quantum-classical classifier in PennyLane that integrates with PyTorch for training on a dataset

Hardware execution

Show how to execute a PennyLane circuit on IBM Quantum hardware using the qiskit.ibmq device with parameter-shift gradients

最佳實務

Start with simulators before deploying to quantum hardware
Use parameter-shift rule for gradient computation on hardware
Initialize parameters carefully to avoid barren plateaus
Cache device objects to reduce initialization overhead

避免

Using backpropagation on quantum hardware (only works on simulators)
Skipping circuit validation on simulators before hardware execution
Using too many qubits beyond simulator capacity
Ignoring noise models when simulating hardware behavior

常見問題

What quantum hardware does PennyLane support?

IBM Quantum, Amazon Braket, Google Cirq, Rigetti, IonQ, and Azure Quantum through device plugins.

What is the maximum number of qubits for simulation?

default.qubit handles about 25 qubits; lightning.qubit extends to about 30 qubits depending on available memory.

How does PennyLane integrate with PyTorch?

Use interface='torch' in QNode decorator; PennyLane tensors are compatible with torch.autograd for backpropagation.

Is my data safe when using cloud quantum hardware?

Circuit definitions are sent to hardware providers; sensitive data should be minimized in circuit parameters.

Why are my gradients vanishing in deep circuits?

Barren plateaus affect deep circuits; try initialization strategies, shallower circuits, or local cost functions.

How does PennyLane compare to Qiskit or Cirq?

PennyLane focuses on variational algorithms and automatic differentiation; Qiskit and Cirq are lower-level frameworks.

開發者詳情

作者

davila7

授權

MIT

儲存庫

https://github.com/davila7/claude-code-templates/tree/main/cli-tool/components/skills/scientific/pennylane

引用

main

檔案結構

📁 references/

📄 advanced_features.md

📄 devices_backends.md

📄 getting_started.md

📄 optimization.md

📄 quantum_chemistry.md

📄 quantum_circuits.md

📄 quantum_ml.md

📄 SKILL.md