Info: Zenodo’s user support line is staffed on regular business days between Dec 23 and Jan 5. Response times may be slightly longer than normal.

Published July 9, 2024 | Version v1
Model Open

Pre-training with Fractional Denoising to Enhance Molecular Property Prediction

  • 1. ROR icon Tsinghua University

Description

 

The pre-trained models described in the paper "Pre-training with Fractional Denoising to Enhance Molecular Property Prediction" are available for various tasks. Below are the details of each pre-trained model:

  1. Pre-trained model for QM9:
    • Filename: step=399999-epoch=8-val_loss=0.1422-test_loss=0.2784-train_per_step=0.1512.ckpt
  2. Pre-trained model for Force Prediction (MD17, MD22, ISO17):
    • Filename: step=105583-epoch=15-val_loss=0.1718-test_loss=0.1715-train_per_step=0.1750.ckpt
  3. Pre-trained model for LBA:
    • Filename: step=199999-epoch=14-val_loss=0.3943-test_loss=0.3931-train_per_step=0.3914.ckpt

Files