Search this site
Embedded Files
  • Home
  • Motivation & Physics
  • Dataset Description
  • Model Architecture
  • Training Strategy
  • Results
  • Code & Reproducibility
 
  • Home
  • Motivation & Physics
  • Dataset Description
  • Model Architecture
  • Training Strategy
  • Results
  • Code & Reproducibility
  • More
    • Home
    • Motivation & Physics
    • Dataset Description
    • Model Architecture
    • Training Strategy
    • Results
    • Code & Reproducibility

Code & Reproducibility

Implementation Details

The entire workflow was implemented in Python using:

  • PyTorch

  • PyTorch Geometric

  • ASE (Atomic Simulation Environment)

  • NumPy and SciPy

Repository Structure

The GitHub repository includes:

  • Dataset loading and graph construction scripts

  • GNN model definition

  • BVSE pretraining pipeline

  • DFT fine-tuning pipeline

  • Scratch baseline training

  • Evaluation and benchmarking scripts

All experiments are reproducible using fixed random seeds and defined train/test splits.

Reproducing Results

To reproduce the results:

  1. Download the LiTraj datasets (nebBVSE122k and nebDFT2k).

  2. Place them in the specified data directory.

  3. Run BVSE pretraining.

  4. Run DFT fine-tuning.

  5. Run evaluation on the held-out DFT test set.

Model checkpoints and evaluation scripts are included.

Google Sites
Report abuse
Page details
Page updated
Google Sites
Report abuse