five

ESM2_AMP

收藏
Figshare2025-06-23 更新2026-04-08 收录
下载链接:
https://figshare.com/articles/dataset/ESM2_AMP/28378157/7
下载链接
链接失效反馈
官方服务:
资源简介:
The feature files for all proteins in the dataset required for the protein-protein interaction prediction framework (i.e., ESM2 embeddings) and the weight files for the models.real_test_dataset_embedding.zip is the feature representation of proteins and protein pair samples in the real_test dataset.ESM2_AMPS.pth is the weight file for the ESM2_AMPS model (Trainning on Pan_dataset);ESM2_AMP_CSE.pth is the weight file for the ESM2_AMP_CSE model;ESM2_DPM.pth is the weight file for the ESM2_DPM model;ESM2_AMPS_Bernett.zip is the weight file for the ESM2_AMPS model (Trainning on Bernett dataset);ESM2_GRU_Bernett.zip is the weight file for the ESM2_GRU model (Trainning on Bernett dataset);ESM2_AMPS and ESM2_AMP_CSE attention_weights.zip are the extracted attention weights when the input dataset is real_test dataset for the ESM2_AMPS and ESM2_AMP_CSE models.segment_model_weights.zip contains feature attribution-related models for "segment 0-9", including the autoencoder model, AE_DNN model, and AE_RF model.cls_segment_eos_model_weights.zip contains feature attribution-related models for "cls_segment 0-9_eos", including the autoencoder model, AE_DNN model, and AE_RF model.10.pan_pairs_features.zip contains cls, segment0-9, and eos features, which are used for training the ESM2_AMPS and ESM2_AMP_CSE models.11.pan_pairs_mean_feature.zip contains ESM2_mean features, which are used for training the ESM2_DPM model.<br>The following are the protein characteristics provided in the three datasets:Pan_dataset_protein_features.npy contains the ESM2 embedding features of all proteins in the Pan_dataset; Bernett_dataset_protein_features.zip is a collection of ESM2 embedding features for all proteins in the Bernett dataset;<br>Multi_species_dataset_protein_features.zip is a collection of ESM2 embedding features for all proteins in the Multi_species_dataset;<br>
提供机构:
Yawen, Sun; rui, wang
创建时间:
2025-06-23
5,000+
优质数据集
54 个
任务类型
进入经典数据集
二维码
社区交流群

面向社区/商业的数据集话题

二维码
科研交流群

面向高校/科研机构的开源数据集话题

数据驱动未来

携手共赢发展

商业合作