Unimol
收藏arXiv2025-09-30 收录
下载链接：
https://bioos-hermite-beijing.tos-cn-beijing.volces.com/unimol_data/pretrain/ligands.tar.gz
下载链接
链接失效反馈官方服务：
资源简介：
该数据集包含了1900万个多样化的分子结构，这些结构被用于预训练分子3D变压器模型。此外，该数据集在脂质纳米颗粒转染效率预测的背景下也得到了应用。其规模达到了1900万个分子结构，任务是对分子3D变压器模型进行预训练。
This dataset contains 19 million diverse molecular structures, which are utilized for pre-training molecular 3D Transformer models. Additionally, this dataset has also been applied in the context of predicting the transfection efficiency of lipid nanoparticles. Its scale reaches 19 million molecular structures, and the task is to pre-train the molecular 3D Transformer model.
提供机构：
Unimol
