search

LirongWu / MAPE-PPI

Code for ICLR 2024 (Spotlight) paper "MAPE-PPI: Towards Effective and Efficient Protein-Protein Interaction Prediction via Microenvironment-Aware Protein Embedding"
MIT License
248 stars 44 forks source link

about share more pretrained ckpt #10

Closed yzf-code closed 2 weeks ago

yzf-code commented 2 weeks ago

Thanks for your great work! I noticed you provide a pre-trained model in ./trained_model/ for PPI prediction on STRING. However, according to your configs, the model structures under different settings of different datasets are not same. Can you share the ckpts of more datasets? For example, the ckpts of SHS27k and SHS148k under bfs and dfs? Thank you very much!

LirongWu commented 2 weeks ago

We are sorry to only provide pre-trained models for STRING because the SHS27k and SHS148k datasets are several orders of magnitude smaller than STRING, and we think that it is not slow to train them from scratch with a suitable combination of hyperparameters.

yzf-code commented 2 weeks ago

yeah, I noticed that, thanks for the reply.