colabfit/aC_JCP_2023_train

aC JCP 2023 train数据集是从216个原子的非晶态部分切割出的训练集，使用VASP软件进行从头计算生成。在计算中使用了LDA交换相关功能以及碳的PAW势。通过熔化-淬火模拟产生了非晶态和液态结构。在多种密度下进行了模拟，以产生多种结构。数据集包含了16,830个数据点和560个钻石结构数据点。此外，还生成了512个原子的系统数据，用于验证较大结构的预测。数据集包含了能量、原子力和Cauchy应力等属性。

The aC JCP 2023 train dataset is a training set cut from the amorphous portion of 216 atoms, generated using ab initio calculations with VASP software. The LDA exchange-correlation functional and the PAW potential for carbon were used in the calculations. Melt-quench simulations were performed to create amorphous and liquid-state structures at various densities. The dataset contains 16,830 data points for the amorphous and liquid states and 560 data points for the diamond structures. Additionally, data for 512-atom systems were generated to validate predictions for larger structures. The dataset includes properties such as energy, atomic forces, and Cauchy stress.