ClemGregArLu/mech_uspto_31k_th

该数据集包含化学反应相关数据，包括反应物、产物的SMILES表示、元素组成、反应索引、步骤索引（正向和逆向）、分割标签（训练/验证/测试）以及用于验证的目标SMILES。数据分为训练集（91805条）、验证集（11396条）和测试集（11625条）。This dataset contains chemical reaction data including SMILES of reactants and products, elemental compositions, reaction indices, step indices (forward and retro), split labels (train/val/test), and target SMILES for validation. The dataset is split into training (91805 examples), validation (11396 examples), and test (11625 examples) sets.