# 📣PaddleMaterials 【PPMat-模型复现】共建计划 ## 任务目标 适配AI+材料领域若干模型,并评测相应的模型在数据集上的效果,保证复现的模型与原始模型精度需基本一致。提交的代码和逻辑需符合repo已有的既定逻辑规范。并将代码合入到github仓库中。 ## 模型列表 | 序号 | 模型名称 | 状态 / PR号 / 认领人 | 材料类型 | 模型结构 | AI4Mat任务 | 备注 | 难度 | 链接 | |-----|-------------|-------------------|----------|----------|-----------------|-------------|----------------|----------------| | 1 | CrystalLLM | | 晶体材料 | Transformer | 结构生成 | | 中等 | https://github.com/lantunes/CrystaLLM | | 2 | wD-MPNN | | 高分子 | GNN | 属性预测 | | 中等 | https://github.com/coleygroup/polymer-chemprop | | 3 | Crystalformer | | 晶体材料 | Transformer | 属性预测 | | 中等 | https://github.com/omron-sinicx/crystalformer | | 4 | NewtonNet | | 小分子 | GNN | 机器学习原子间势函数 | | 中等 | https://github.com/THGLab/NewtonNet | | 5 | ECDFormer | | 小分子 | GNN | 谱图解析 | | 复杂 | https://github.com/HowardLi1984/ECDFormer | | 6 | TrinityLLM | | 高分子 | LLM | 属性预测 | 高优🌟 | 中等 | https://github.com/ningliu-iga/TrinityLLM | | 7 | AlloyGAN | | 无机材料 | LLM | 工作流 | | 中等 | https://github.com/photon-git/AlloyGAN | | 8 | GDI-NN | | 小分子 | GNN-PINN | 热力学计算 | 高优🌟 | 中等 | https://git.rwth-aachen.de/avt-svt/public/GDI-NN | | 9 | SFIN | <img src="https://img.shields.io/badge/状态-进行中-2ECC71" /> #244 @ADream-ki <br> | 晶体材料 | CNN | 谱图增强 | | 复杂 | https://github.com/HeasonLee/SFIN | |10 | DM2 | | 非晶无机材料 | 扩散模型 | 结构生成 | 高优🌟 | 复杂 | https://github.com/digital-synthesis-lab/DM2 | |11 | DiffSyn | | 无机材料 | 扩散模型 | 合成规划 | 高优🌟 | 复杂 | https://github.com/eltonpan/zeosyn_gen | |12 | MOFDiff | | MOF | 扩散模型 | 结构生成 | | 复杂 | https://github.com/microsoft/MOFDiff | |13 | SchNet | | 小分子 | GNN | 属性预测 | 高优🌟 | 中等 | https://github.com/atomistic-machine-learning/SchNet | |14 | SphereNet | | 小分子 | GNN | 属性预测 | 高优🌟 | 复杂 | https://github.com/divelab/DIG | |15 | GPWNO | ⛔ | 晶体材料&小分子 | 算子学习 | 机器学习电子结构 | **MIIT-1** | | https://github.com/seongsukim-ml/GPWNO | |16 | UMA | ⛔ | 多类型 | GNN | 机器学习原子间势函数&属性预测 | **MIIT-1** | | https://github.com/facebookresearch/fairchem/tree/main/src/fairchem/core/models/uma | |17 | DeepH / DeepH-E3 | ⛔ | 多类型 | GNN | 密度泛函 | **MIIT-1** | | https://github.com/mzjb/DeepH-pack | |18 | TransPolymer | ⛔ | 高分子 | Transformer | 属性预测 | **MIIT-2** | | https://github.com/ChangwenXu98/TransPolymer | |19 | DenseGNN | ⛔ | 无机材料&小分子 | GNN | 属性预测 | **MIIT-2** | | https://github.com/dhw059/DenseGNN | |20 | liflow | ⛔ | 晶体材料 | 流匹配 | 分子动力学模拟 | **MIIT-2** | | https://github.com/learningmatter-mit/liflow | |21 | GMTNet | ⛔ | 无机材料 | GNN | 属性预测 | **MIIT-2** | | https://github.com/divelab/AIRS/tree/main/OpenMat/GMTNet | |22 | GE-GNN | ⛔ | 小分子溶液体系 | GNN-PINN | 热力学 | **MIIT-2** | | https://git.rwth-aachen.de/avt-svt/public/GDI-NN | |23 | RXNGraphormer | ⛔ | 小分子 | GNN | 化学合成 | **MIIT-3** | | https://github.com/licheng-xu-echo/RXNGraphormer | |24 | PXRDGen | ⛔ | 晶体材料 | 多模态 | 谱图解析 | **MIIT-3** | | https://codeocean.com/capsule/7727770/tree/v1 | |25 | SGEquiDiff | ⛔ | 晶体材料 | Diffusion | 结构生成 | **MIIT-3** | | https://github.com/rees-c/sgequidiff | |26 | Chemeleon2 | <img src="https://img.shields.io/badge/状态-进行中-2ECC71" /> #243 @learncat163 <br> | 无机材料 | RL+VAE/DM | 结构生成 | **MIIT-3** | | https://github.com/hspark1212/chemeleon2 | |27 | MatInvent | ⛔ | 无机材料 | RL+DM | 结构生成 | **MIIT-3** | | https://github.com/schwallergroup/matinvent | |28 | Uni-Mol | | 小分子 | GNN | 属性预测 | 高优🌟 | 中等 | https://github.com/deepmodeling/Uni-Mol/tree/main/unimol2 |
📣PaddleMaterials 【PPMat-模型复现】共建计划
任务目标
适配AI+材料领域若干模型,并评测相应的模型在数据集上的效果,保证复现的模型与原始模型精度需基本一致。提交的代码和逻辑需符合repo已有的既定逻辑规范。并将代码合入到github仓库中。
模型列表