vgon 0.0.0__tar.gz

vgon-0.0.0/PKG-INFO

@@ -0,0 +1,29 @@
+Metadata-Version: 2.4
+Name: vgon
+Version: 0.0.0
+Summary: Variational Generative Optimization Network (VGON).
+Author: Lingxia Zhang, Xiaodie Lin
+Project-URL: Homepage, https://vgon.funqitang.cn
+Project-URL: Repository, https://github.com/zhangjianjianzz/VGON.git
+Project-URL: Documentation, https://vgon.funqitang.cn
+Requires-Python: >=3.11
+Provides-Extra: core
+Requires-Dist: torch; extra == "core"
+Requires-Dist: toqito>=1.1.2; extra == "core"
+Requires-Dist: cvxpy>=1.7.2; extra == "core"
+Requires-Dist: cvxpylayers>=0.1.9; extra == "core"
+Requires-Dist: pennylane>=0.42.3; extra == "core"
+Requires-Dist: pennylane-lightning>=0.42.0; extra == "core"
+Provides-Extra: plot
+Requires-Dist: pandas>=2.3.2; extra == "plot"
+Requires-Dist: ptiprince>=0.2.7; extra == "plot"
+Requires-Dist: scipy>=1.16.2; extra == "plot"
+Requires-Dist: seaborn>=0.13.2; extra == "plot"
+Provides-Extra: notebook
+Requires-Dist: ipykernel>=6.30.1; extra == "notebook"
+Provides-Extra: doc
+Requires-Dist: mkdocs-material; extra == "doc"
+Requires-Dist: mkdocs-literate-nav>=0.6.2; extra == "doc"
+Requires-Dist: mdx-truly-sane-lists>=1.3; extra == "doc"
+Provides-Extra: full
+Requires-Dist: vgon[core,doc,notebook,plot]; extra == "full"

vgon-0.0.0/README.md

@@ -0,0 +1,178 @@
+# VGON Environment Setup
+
+VGON scripts here require Python version 3.11 and above.
+
+This guide provides instructions for setting up the VGON (Variational Generative Optimization Network) environment using either `uv` or `pip`.
+
+## Prerequisites
+
+- Python 3.11 or higher
+- Git
+
+## Method 1: Using uv (Recommended)
+
+[uv](https://github.com/astral-sh/uv) is a fast Python package manager and project manager, whose installation is simple as [guide](https://docs.astral.sh/uv/#installation)
+
+
+### Setup with uv
+1. Clone the repository to local disk.
+    ```bash
+    git clone https://github.com/zhangjianjianzz/VGON.git && cd VGON
+    ```
+2. Create virtual environment if it wasn't contained
+    ```bash
+    uv venv .venv
+    ```
+    The python virtual environment will be located at `.venv` fold under root path in default. Linux/macOS users could activate it in shell by typing
+    ```bash
+    source .venv/bin/activate
+    ```
+3. Install dependecies in `.venv` for VGON scripts 
+    ```bash
+    uv sync --extra full
+    ```
+    `plot` and `notebook` groups are optional for VGON itself. 
+
+4. Running experiments with uv
+    ```bash
+    # Run specific experiments
+    uv run python BP/HXXZ/vgon_xxz.py
+    uv run python Gap/Mix.py
+    uv run python Degeneracy/H232/H232.py
+
+    # Or activate environment first
+    source .venv/bin/activate
+    python BP/HXXZ/plot.py
+    ```
+
+## Method 2: Using pip
+
+### Setup with pip
+
+```bash
+# Clone the repository
+git clone <your-repo-url>
+cd VGON
+
+# Create virtual environment
+python -m venv venv
+
+# Activate virtual environment
+source venv/bin/activate  # On Linux/macOS
+# or
+venv\Scripts\activate     # On Windows
+
+# Upgrade pip
+pip install --upgrade pip
+
+# Install PyTorch (CPU version)
+pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cpu
+
+# For GPU support (CUDA 12.1), use instead:
+# pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu121
+
+# Install the package with dependencies
+pip install -e .
+
+# Install optional dependencies
+pip install -e ".[plot]"     # For plotting
+pip install -e ".[dev]"      # For development
+pip install -e ".[gpu]"      # For GPU acceleration
+pip install -e ".[all]"      # Install everything
+
+# Manual installation of ptitprince (if needed)
+pip install git+https://github.com/pog87/PtitPrince.git
+```
+
+## Verify Installation
+
+Test your installation by running a simple example:
+
+```python
+import torch
+import pennylane as qml
+import numpy as np
+
+# Test basic functionality
+print(f"PyTorch version: {torch.__version__}")
+print(f"PennyLane version: {qml.__version__}")
+print(f"CUDA available: {torch.cuda.is_available()}")
+
+# Test quantum device
+dev = qml.device("default.qubit", wires=2)
+print("✓ PennyLane quantum device created successfully")
+```
+
+## Project Structure
+
+```
+VGON/
+├── BP/              # Barren Plateau experiments
+│   ├── HXXZ/        # Heisenberg XXZ model
+│   └── Z1Z2/        # Z1Z2 model
+├── Degeneracy/      # Degeneracy detection experiments
+│   ├── H232/        # H232 Hamiltonian
+│   └── MG/          # Graph states
+├── Gap/             # Nonlocality gap experiments
+└── results/         # Experimental results
+```
+
+## Running Experiments
+
+### Barren Plateau Experiments
+```bash
+# HXXZ model
+python BP/HXXZ/vgon_xxz.py    # Run VGON training
+python BP/HXXZ/vqe_xxz.py     # Run VQE baseline
+python BP/HXXZ/plot.py        # Generate plots
+
+# Z1Z2 model
+python BP/Z1Z2/vgon_z1z2.py
+python BP/Z1Z2/vqe_z1z2.py
+python BP/Z1Z2/plot.py
+```
+
+### Degeneracy Detection
+```bash
+# H232 Hamiltonian
+python Degeneracy/H232/H232.py
+python Degeneracy/H232/plot.py
+
+# Graph states
+python Degeneracy/MG/MG.py
+python Degeneracy/MG/plot.py
+```
+
+### Nonlocality Gap
+```bash
+python Gap/Mix.py              # Train gap model
+matlab -batch "run('Gap/plotGap.m')"  # Generate MATLAB plots
+```
+
+## Troubleshooting
+
+### Common Issues
+
+1. **CUDA not found**: If you have a GPU but CUDA is not detected, reinstall PyTorch with CUDA support
+2. **PennyLane device errors**: Ensure you have the correct PennyLane plugins installed
+3. **Memory issues**: Reduce batch sizes in the configuration files
+4. **Import errors**: Make sure all dependencies are installed correctly
+
+### GPU Setup
+
+For GPU acceleration:
+
+
+### Development Setup
+
+For contributors:
+
+## Support
+
+- Paper: [Commun Phys 8, 334 (2025)](https://doi.org/10.1038/s42005-025-02261-4)
+- Issues: Create an issue on the GitHub repository
+- Documentation: See individual module docstrings and comments
+
+# [Development Containers](https://containers.dev/)
+DevContainers' configuration in vscode/GitHub Codespace.
+

vgon-0.0.0/pyproject.toml

@@ -0,0 +1,95 @@
+[build-system]
+requires = ["setuptools>=65.5.0", "wheel>=0.38.0"]
+build-backend = "setuptools.build_meta"
+
+[project]
+name = "vgon"
+dynamic = ["version"]
+description = "Variational Generative Optimization Network (VGON)."
+authors = [
+    {name = "Lingxia Zhang"},
+    {name = "Xiaodie Lin"}
+]
+requires-python = ">=3.11"
+
+dependencies = []
+
+[project.optional-dependencies]
+core = [
+    "torch",
+    "toqito>=1.1.2",
+    "cvxpy>=1.7.2",
+    "cvxpylayers>=0.1.9",
+    "pennylane>=0.42.3",
+    "pennylane-lightning>=0.42.0",
+]
+
+# uv sync --extra plot --extra notebook -U
+# you could cancle 'plot' block to use basic feature.
+plot = [
+    "pandas>=2.3.2",
+    "ptiprince>=0.2.7",
+    "scipy>=1.16.2",
+    "seaborn>=0.13.2",
+]
+notebook = [
+    "ipykernel>=6.30.1",
+]
+doc = [
+    "mkdocs-material",
+    "mkdocs-literate-nav>=0.6.2",
+    "mdx-truly-sane-lists>=1.3",
+]
+
+# All optional dependencies
+full = [
+    "vgon[core,plot,notebook,doc]"
+]
+
+
+# package-location
+[tool.setuptools]
+package-dir = {"" = "src"}
+include-package-data = true
+# package-name-rule
+[tool.setuptools.packages.find]
+where = ["src"]
+include = ["vgon", "vgon.*"]
+exclude = ["tests", "docs"]
+# package-data-rule
+[tool.setuptools.package-data]
+vgon = ["Examples/**/*.mat", "Examples/**/*.csv", "Examples/**/*.json"]
+[tools.setuptools_scm]
+write_to = "src/vgon/_version.py"
+version_scheme = "guess-next-dev"
+local_scheme = "no-local-version"
+
+[project.urls]
+Homepage = "https://vgon.funqitang.cn"
+Repository = "https://github.com/zhangjianjianzz/VGON.git"
+Documentation = "https://vgon.funqitang.cn"
+
+
+
+[tool.uv.sources]
+# "pytorch-cpu", "pytorch-cuda"
+# For users outside of CUDA eco, redirect torch.index to "pytorch-cpu"
+torch = { index = "pytorch-cuda" }
+# dependency-updated repo forked from [pog87/PtitPrince](https://github.com/pog87/PtitPrince.git) 
+ptiprince = {git = "https://github.com/leo2www/PtitPrince.git"}
+mkdocs-material = { git = "https://github.com/squidfunk/mkdocs-material.git" }
+
+# pennylane = { git = "https://github.com/PennyLaneAI/pennylane.git" }
+# seaborn = {git = "https://github.com/mwaskom/seaborn.git"}
+
+
+
+[[tool.uv.index]]
+name = "pytorch-cpu"
+url = "https://download.pytorch.org/whl/cpu"
+explicit = true
+
+[[tool.uv.index]]
+name = "pytorch-cuda"
+url = "https://download.pytorch.org/whl/cu129"
+explicit = true

vgon-0.0.0/setup.cfg

@@ -0,0 +1,4 @@
+[egg_info]
+tag_build = 
+tag_date = 0
+

vgon-0.0.0/src/vgon/__init__.py

@@ -0,0 +1 @@
+from vgon._version import __version__

vgon-0.0.0/src/vgon/examples/gap/__init__.py

@@ -0,0 +1,9 @@
+from .lo import E2Lo
+from .locc import E2Locc
+from .ns import __initialize_Ns__, Ns
+
+__all__ = {
+    "__initialize_Ns__",
+    "E2Lo",
+    "E2Locc"
+}

vgon-0.0.0/src/vgon/examples/gap/lo.py

@@ -0,0 +1,46 @@
+import cvxpy as cp
+import numpy as np
+
+from .ns import __initialize_Ns__
+
+
+def E2Lo():
+    rho = cp.Parameter((9, 9))
+    e1 = cp.Parameter(1)
+
+    # parameters
+    d = X = A = 3
+    size = d**2
+
+    # formulate sdp
+    # variables
+    P_lo = cp.Variable((2, X**2 * A**2), nonneg=True)  # [gamma, xyab]
+    M0_lo = cp.Variable((size, size), hermitian=True)
+    M1_lo = cp.Variable((size, size), hermitian=True)
+    e2_lo = cp.Variable(1, nonneg=True)
+    # construct MC, MU
+    Ns = __initialize_Ns__()
+
+    # reshape in cvxpy rearranges elements column by column, so transpose (.T) is needed
+    PU_lo = cp.reshape(P_lo[0], (1, size**2))
+    MU_lo = cp.reshape(PU_lo @ Ns, (size, size)).T
+    PC_lo = cp.reshape(P_lo[1], (1, size**2))
+    MC_lo = cp.reshape(PC_lo @ Ns, (size, size)).T
+
+    # constraints
+    constraints = [M0_lo >> 0, M1_lo >> 0]
+    constraints += [
+        e1 * np.eye(size, dtype=complex) - MC_lo == M0_lo + cp.partial_transpose(M1_lo, dims=(d, d), axis=1)
+    ]
+    # constraints of P - lo
+    Pxy_ab_lo = cp.reshape(cp.sum(P_lo, axis=0), (A**2, X**2)).T
+    for i in range(Pxy_ab_lo.shape[1] - 1):
+        for j in range(i + 1, Pxy_ab_lo.shape[1]):
+            constraints += [Pxy_ab_lo[:, i] - Pxy_ab_lo[:, j] == np.zeros(Pxy_ab_lo.shape[0])]
+    constraints += [cp.sum(Pxy_ab_lo[:, 0]) == 1]
+    constraints += [e2_lo == cp.real(cp.trace(MU_lo @ rho))]
+
+    # problem
+    parameter = [rho, e1]
+    variable = [e2_lo, P_lo, M0_lo, M1_lo]
+    return cp.Problem(cp.Minimize(cp.real(cp.trace(MU_lo @ rho))), constraints), parameter, variable

vgon-0.0.0/src/vgon/examples/gap/locc.py

@@ -0,0 +1,57 @@
+import cvxpy as cp
+import numpy as np
+
+from .ns import __initialize_Ns__
+
+
+def E2Locc():
+    rho = cp.Parameter((9, 9))
+    e1 = cp.Parameter(1)
+
+    # parameters
+    d = X = A = 3
+    size = d**2
+
+    # formulate sdp
+    # variables
+    P_locc = cp.Variable((2, X**2 * A**2), nonneg=True)  # [gamma, xyab]
+    M0_locc = cp.Variable((size, size), hermitian=True)
+    M1_locc = cp.Variable((size, size), hermitian=True)
+    e2_locc = cp.Variable(1, nonneg=True)
+    # construct MC, MU
+    Ns = __initialize_Ns__()
+
+
+    # reshape in cvxpy rearranges elements clolumn by clolumn, so transpose (.T) is needed
+    PU_locc = cp.reshape(P_locc[0], (1, size**2))
+    MU_locc = cp.reshape(PU_locc @ Ns, (size, size)).T
+    PC_locc = cp.reshape(P_locc[1], (1, size**2))
+    MC_locc = cp.reshape(PC_locc @ Ns, (size, size)).T
+
+    # constraints
+    constraints = [M0_locc >> 0, M1_locc >> 0]
+    constraints += [
+        e1 * np.eye(size, dtype=complex) - MC_locc == M0_locc + cp.partial_transpose(M1_locc, dims=(d, d), axis=1)
+    ]
+    # constraints of P - locc
+    Pxya_b_locc = cp.reshape(cp.sum(P_locc, axis=0), (A, X**2 * A)).T
+    constraints += [
+        Pxya_b_locc[:, 0] - Pxya_b_locc[:, 1] == np.zeros(Pxya_b_locc.shape[0]),
+        Pxya_b_locc[:, 0] - Pxya_b_locc[:, 2] == np.zeros(Pxya_b_locc.shape[0])
+    ]
+    P_xy_a_locc = cp.reshape(Pxya_b_locc[:, 0], (A, X**2)).T
+    sum_matrix = np.array([[1, 1, 1, 0, 0, 0, 0, 0, 0], [0, 0, 0, 1, 1, 1, 0, 0, 0], [0, 0, 0, 0, 0, 0, 1, 1, 1]])
+    P_x_a_locc = sum_matrix @ P_xy_a_locc
+    constraints += [
+        P_x_a_locc[:, 0] - P_x_a_locc[:, 1] == np.zeros(P_x_a_locc.shape[0]),
+        P_x_a_locc[:, 0] - P_x_a_locc[:, 2] == np.zeros(P_x_a_locc.shape[0])
+    ]
+    constraints += [cp.sum(P_x_a_locc[:, 0]) == 1]
+    constraints += [e2_locc == cp.real(cp.trace(MU_locc @ rho))]
+
+    # problem
+    parameter = [rho, e1]
+    variable = [e2_locc, P_locc, M0_locc, M1_locc]
+    return cp.Problem(cp.Minimize(cp.real(cp.trace(MU_locc @ rho))), constraints), parameter, variable
+
+

vgon-0.0.0/src/vgon/examples/gap/ns.py

@@ -0,0 +1,44 @@
+import numpy as np
+
+
+Ns = None  # [xyab, N_xyab.reshape(-1)]
+
+
+def __initialize_Ns__():
+    global Ns
+    if Ns is None:
+        # parameters
+        X = A = d = 3
+        # define measurements
+        N_xa = [[] for _ in range(X)]  # [x][a]
+        # N_xa[0]
+        vec = np.array([[1], [0], [0]], dtype=complex)
+        N_xa[0].append(vec @ vec.conj().T)
+        vec = np.array([[0], [1], [0]], dtype=complex)
+        N_xa[0].append(vec @ vec.conj().T)
+        vec = np.array([[0], [0], [1]], dtype=complex)
+        N_xa[0].append(vec @ vec.conj().T)
+        # N_xa[1]
+        vec = np.array([[np.exp(1j * 2 * np.pi / 3)], [np.exp(-1j * 2 * np.pi / 3)], [1]], dtype=complex) / np.sqrt(3)
+        N_xa[1].append(vec @ vec.conj().T)
+        vec = np.array([[np.exp(-1j * 2 * np.pi / 3)], [np.exp(1j * 2 * np.pi / 3)], [1]], dtype=complex) / np.sqrt(3)
+        N_xa[1].append(vec @ vec.conj().T)
+        vec = np.array([[1], [1], [1]], dtype=complex) / np.sqrt(3)
+        N_xa[1].append(vec @ vec.conj().T)
+        # N_xa[2]
+        vec = np.array([[1], [-1], [0]], dtype=complex) / (-np.sqrt(2))
+        N_xa[2].append(vec @ vec.conj().T)
+        vec = np.array([[0], [0], [1]], dtype=complex)
+        N_xa[2].append(vec @ vec.conj().T)
+        vec = np.array([[1], [1], [0]], dtype=complex) / (-np.sqrt(2))
+        N_xa[2].append(vec @ vec.conj().T)
+
+        # calculate final measurement matrix
+        Ns = np.zeros((X**2 * A**2, d**4), dtype=complex)  # [xyab, N_xyab.reshape(-1)]
+        for x in range(X):
+            for y in range(X):
+                for a in range(A):
+                    for b in range(A):
+                        Ns[x * (X * A**2) + y * A**2 + a * A + b, :] = np.kron(N_xa[x][a], N_xa[y][b]).reshape(-1)
+
+    return Ns

vgon-0.0.0/src/vgon.egg-info/PKG-INFO

@@ -0,0 +1,29 @@
+Metadata-Version: 2.4
+Name: vgon
+Version: 0.0.0
+Summary: Variational Generative Optimization Network (VGON).
+Author: Lingxia Zhang, Xiaodie Lin
+Project-URL: Homepage, https://vgon.funqitang.cn
+Project-URL: Repository, https://github.com/zhangjianjianzz/VGON.git
+Project-URL: Documentation, https://vgon.funqitang.cn
+Requires-Python: >=3.11
+Provides-Extra: core
+Requires-Dist: torch; extra == "core"
+Requires-Dist: toqito>=1.1.2; extra == "core"
+Requires-Dist: cvxpy>=1.7.2; extra == "core"
+Requires-Dist: cvxpylayers>=0.1.9; extra == "core"
+Requires-Dist: pennylane>=0.42.3; extra == "core"
+Requires-Dist: pennylane-lightning>=0.42.0; extra == "core"
+Provides-Extra: plot
+Requires-Dist: pandas>=2.3.2; extra == "plot"
+Requires-Dist: ptiprince>=0.2.7; extra == "plot"
+Requires-Dist: scipy>=1.16.2; extra == "plot"
+Requires-Dist: seaborn>=0.13.2; extra == "plot"
+Provides-Extra: notebook
+Requires-Dist: ipykernel>=6.30.1; extra == "notebook"
+Provides-Extra: doc
+Requires-Dist: mkdocs-material; extra == "doc"
+Requires-Dist: mkdocs-literate-nav>=0.6.2; extra == "doc"
+Requires-Dist: mdx-truly-sane-lists>=1.3; extra == "doc"
+Provides-Extra: full
+Requires-Dist: vgon[core,doc,notebook,plot]; extra == "full"

vgon-0.0.0/src/vgon.egg-info/SOURCES.txt

@@ -0,0 +1,13 @@
+README.md
+pyproject.toml
+src/vgon/__init__.py
+src/vgon.egg-info/PKG-INFO
+src/vgon.egg-info/SOURCES.txt
+src/vgon.egg-info/dependency_links.txt
+src/vgon.egg-info/requires.txt
+src/vgon.egg-info/top_level.txt
+src/vgon/examples/__init__.py
+src/vgon/examples/gap/__init__.py
+src/vgon/examples/gap/lo.py
+src/vgon/examples/gap/locc.py
+src/vgon/examples/gap/ns.py

vgon-0.0.0/src/vgon.egg-info/dependency_links.txt

@@ -0,0 +1 @@
+

vgon-0.0.0/src/vgon.egg-info/requires.txt

@@ -0,0 +1,25 @@
+
+[core]
+torch
+toqito>=1.1.2
+cvxpy>=1.7.2
+cvxpylayers>=0.1.9
+pennylane>=0.42.3
+pennylane-lightning>=0.42.0
+
+[doc]
+mkdocs-material
+mkdocs-literate-nav>=0.6.2
+mdx-truly-sane-lists>=1.3
+
+[full]
+vgon[core,doc,notebook,plot]
+
+[notebook]
+ipykernel>=6.30.1
+
+[plot]
+pandas>=2.3.2
+ptiprince>=0.2.7
+scipy>=1.16.2
+seaborn>=0.13.2

vgon-0.0.0/src/vgon.egg-info/top_level.txt

@@ -0,0 +1 @@
+vgon