DeepHall-data

# Data for the Study: "Taming Landau Level Mixing in Fractional Quantum Hall States with Deep Learning"

This dataset contains the raw data and plotting code used in the study titled:

"[Taming Landau Level Mixing in Fractional Quantum Hall States with Deep Learning](https://arxiv.org/abs/2412.14795)"

The relevant code is available at https://github.com/bytedance/DeepHall.

## Data Files

The data directories are named following the convention `n{number of electrons}l{flux}/k{interaction strength kappa}`.

For each calculation, the following data are included:

- `config.yml`: Hyperparameters used in the simulation.
- `train_stats.csv`: Energy, angular momentum, and other quantities at each training step.
- `ckpt_xxxxxx.npz`: Checkpoint of the final state, including network parameters, walker coordinates, and other relevant quantities.
- `energy/`: NetObs energy evaluation results with fixed parameters.
- `1rdm/`, `overlap/`, `pair_corr/`: NetObs evaluation of the one-body reduced density matrix (1-RDM), overlap with the Laughlin wavefunction, and the pair correlation function.

For the electron densities of quasiparticle and quasihole states, the data are placed in `nxlx/kx/lz{angular momentum in the z direction}`, with:

- `density/`: NetObs evaluation of electron density.

## Code for Plotting Figures

- `prep.py`: Script to process data and create `energy_vs_n.csv` and `llm_1_3.csv` for use in plotting.
- `plot.ipynb`: Jupyter notebook for generating the plots.

# 本研究配套数据集：《利用深度学习（Deep Learning）调控分数量子霍尔态（Fractional Quantum Hall States）中的朗道能级混合（Landau Level Mixing）》
本数据集包含题为上述标题的研究（https://arxiv.org/abs/2412.14795）中所用的原始数据与绘图代码。相关代码可在https://github.com/bytedance/DeepHall获取。

## 数据文件
数据目录遵循`n{电子数}l{磁通数}/k{相互作用强度κ}`的命名规范。
针对每一项计算，所包含的数据如下：
- `config.yml`：模拟过程中使用的超参数（hyperparameters）
- `train_stats.csv`：各训练步数对应的能量、角动量及其他相关物理量
- `ckpt_xxxxxx.npz`：最终状态的检查点文件，包含网络参数、行走者坐标及其他相关物理量
- `energy/`：固定参数下的NetObs能量评估结果
- `1rdm/`、`overlap/`、`pair_corr/`：分别为一体约化密度矩阵（one-body reduced density matrix, 1-RDM）、与劳克林波函数（Laughlin wavefunction）的重叠度、对关联函数（pair correlation function）的NetObs评估结果。

针对准粒子（quasiparticle）与准空穴（quasihole）态的电子密度，数据存储于`nxlx/kx/lz{z方向角动量}`目录下，其中：
- `density/`：电子密度的NetObs评估结果。

## 绘图代码
- `prep.py`：用于处理数据并生成绘图所需的`energy_vs_n.csv`与`llm_1_3.csv`的脚本
- `plot.ipynb`：用于生成各类绘图的Jupyter Notebook。